diff --git a/Cargo.lock b/Cargo.lock
index 99f5f48d2..5294e442f 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -310,17 +310,6 @@ version = "1.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c08606f8c3cbf4ce6ec8e28fb0014a2c086708fe954eaa885384a6165172e7e8"
 
-[[package]]
-name = "automod"
-version = "1.0.16"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e8b5778837666541195063243828c5b6139221b47dc4ec3ba81738e532469ab1"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn 2.0.117",
-]
-
 [[package]]
 name = "aws-lc-rs"
 version = "1.16.1"
@@ -729,15 +718,6 @@ version = "0.9.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c2459377285ad874054d797f3ccebf984978aa39129f6eafde5cdc8315b612f8"
 
-[[package]]
-name = "content_inspector"
-version = "0.2.4"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b7bda66e858c683005a53a9a60c69a4aca7eeaa45d124526e389f7aec8e62f38"
-dependencies = [
- "memchr",
-]
-
 [[package]]
 name = "convert_case"
 version = "0.10.0"
@@ -1493,6 +1473,7 @@ dependencies = [
  "fabro-server",
  "fabro-store",
  "fabro-telemetry",
+ "fabro-test",
  "fabro-types",
  "fabro-util",
  "fabro-validate",
@@ -1526,7 +1507,6 @@ dependencies = [
  "tracing",
  "tracing-appender",
  "tracing-subscriber",
- "trycmd",
  "ulid",
  "walkdir",
  "x509-parser",
@@ -1926,6 +1906,16 @@ dependencies = [
  "uuid",
 ]
 
+[[package]]
+name = "fabro-test"
+version = "0.176.2"
+dependencies = [
+ "assert_cmd",
+ "insta",
+ "regex",
+ "tempfile",
+]
+
 [[package]]
 name = "fabro-tracker"
 version = "0.176.2"
@@ -2776,16 +2766,6 @@ version = "2.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "135b12329e5e3ce057a9f972339ea52bc954fe1e9358ef27f95e89716fbc5424"
 
-[[package]]
-name = "humantime-serde"
-version = "1.1.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "57a3db5ea5923d99402c94e9feb261dc5ee9b4efa158b0315f788cf549cc200c"
-dependencies = [
- "humantime",
- "serde",
-]
-
 [[package]]
 name = "hyper"
 version = "1.8.1"
@@ -3086,6 +3066,7 @@ checksum = "e82db8c87c7f1ccecb34ce0c24399b8a73081427f3c7c50a5d597925356115e4"
 dependencies = [
  "console 0.15.11",
  "once_cell",
+ "regex",
  "similar",
  "tempfile",
 ]
@@ -4134,16 +4115,6 @@ dependencies = [
  "windows-sys 0.61.2",
 ]
 
-[[package]]
-name = "os_pipe"
-version = "1.2.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7d8fae84b431384b68627d0f9b3b1245fcf9f46f6c0e3dc902e9dce64edd1967"
-dependencies = [
- "libc",
- "windows-sys 0.61.2",
-]
-
 [[package]]
 name = "ouroboros"
 version = "0.18.5"
@@ -4693,26 +4664,6 @@ dependencies = [
  "rand_core 0.9.5",
 ]
 
-[[package]]
-name = "rayon"
-version = "1.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "368f01d005bf8fd9b1206fb6fa653e6c4a81ceb1466406b81792d87c5677a58f"
-dependencies = [
- "either",
- "rayon-core",
-]
-
-[[package]]
-name = "rayon-core"
-version = "1.13.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "22e18b0f0062d30d4230b2e85ff77fdfe4326feb054b9783a3460d8435c8ab91"
-dependencies = [
- "crossbeam-deque",
- "crossbeam-utils",
-]
-
 [[package]]
 name = "redox_syscall"
 version = "0.5.18"
@@ -5641,37 +5592,6 @@ dependencies = [
  "serde",
 ]
 
-[[package]]
-name = "snapbox"
-version = "0.6.24"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6c1abc378119f77310836665f8523018532cf7e3faeb3b10b01da5a7321bf8e1"
-dependencies = [
- "anstream",
- "anstyle",
- "content_inspector",
- "dunce",
- "filetime",
- "libc",
- "normalize-line-endings",
- "os_pipe",
- "similar",
- "snapbox-macros",
- "tempfile",
- "wait-timeout",
- "walkdir",
- "windows-sys 0.60.2",
-]
-
-[[package]]
-name = "snapbox-macros"
-version = "0.4.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3b750c344002d7cc69afb9da00ebd9b5c0f8ac2eb7d115d9d45d5b5f47718d74"
-dependencies = [
- "anstream",
-]
-
 [[package]]
 name = "socket2"
 version = "0.6.2"
@@ -6341,7 +6261,7 @@ dependencies = [
  "serde",
  "serde_spanned 0.6.9",
  "toml_datetime 0.6.11",
- "toml_edit 0.22.27",
+ "toml_edit",
 ]
 
 [[package]]
@@ -6391,21 +6311,6 @@ dependencies = [
  "winnow 0.7.14",
 ]
 
-[[package]]
-name = "toml_edit"
-version = "0.23.10+spec-1.0.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "84c8b9f757e028cee9fa244aea147aab2a9ec09d5325a9b01e0a49730c2b5269"
-dependencies = [
- "indexmap 2.13.0",
- "serde_core",
- "serde_spanned 1.0.4",
- "toml_datetime 0.7.5+spec-1.1.0",
- "toml_parser",
- "toml_writer",
- "winnow 0.7.14",
-]
-
 [[package]]
 name = "toml_parser"
 version = "1.0.9+spec-1.1.0"
@@ -6553,24 +6458,6 @@ version = "0.2.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e421abadd41a4225275504ea4d6566923418b7f05506fbc9c0fe86ba7396114b"
 
-[[package]]
-name = "trycmd"
-version = "0.15.11"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a81ea3136ddc88e19c2cc2eb3176b72abee4e831367cd8949f2a88ac5497e64e"
-dependencies = [
- "anstream",
- "automod",
- "glob",
- "humantime",
- "humantime-serde",
- "rayon",
- "serde",
- "shlex",
- "snapbox",
- "toml_edit 0.23.10+spec-1.0.0",
-]
-
 [[package]]
 name = "tungstenite"
 version = "0.26.2"
diff --git a/Cargo.toml b/Cargo.toml
index 582bb8e29..ed19abfab 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -55,6 +55,7 @@ hmac = "0.12"
 sha2 = "0.10"
 hex = "0.4"
 insta = "1"
+fabro-test = { path = "lib/crates/fabro-test" }
 tokio-tungstenite = { version = "0.26", features = ["rustls-tls-webpki-roots"] }
 futures-util = "0.3"
 openssh = "0.11"
diff --git a/lib/crates/fabro-cli/Cargo.toml b/lib/crates/fabro-cli/Cargo.toml
index 82c1f652a..6b1521592 100644
--- a/lib/crates/fabro-cli/Cargo.toml
+++ b/lib/crates/fabro-cli/Cargo.toml
@@ -99,4 +99,4 @@ predicates = "3"
 serde_json.workspace = true
 tempfile = "3"
 httpmock = "0.8"
-trycmd = "0.15"
+fabro-test = { workspace = true }
diff --git a/lib/crates/fabro-cli/tests/cmd/config/help.trycmd b/lib/crates/fabro-cli/tests/cmd/config/help.trycmd
deleted file mode 100644
index b9f098c65..000000000
--- a/lib/crates/fabro-cli/tests/cmd/config/help.trycmd
+++ /dev/null
@@ -1,9 +0,0 @@
-```console
-$ fabro config --help
-[..]
-...
-
-$ fabro config show --help
-[..]
-...
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/cp/help.trycmd b/lib/crates/fabro-cli/tests/cmd/cp/help.trycmd
deleted file mode 100644
index 192511a98..000000000
--- a/lib/crates/fabro-cli/tests/cmd/cp/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro cp --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/doctor/dry-run-flag.stdout b/lib/crates/fabro-cli/tests/cmd/doctor/dry-run-flag.stdout
deleted file mode 100644
index 50fe30806..000000000
--- a/lib/crates/fabro-cli/tests/cmd/doctor/dry-run-flag.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]Fabro Doctor[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/doctor/dry-run-flag.toml b/lib/crates/fabro-cli/tests/cmd/doctor/dry-run-flag.toml
deleted file mode 100644
index 1cdddd1d8..000000000
--- a/lib/crates/fabro-cli/tests/cmd/doctor/dry-run-flag.toml
+++ /dev/null
@@ -1,10 +0,0 @@
-bin.name = "fabro"
-args = ["doctor", "--dry-run"]
-
-[env]
-inherit = false
-
-[env.add]
-PATH = "/usr/local/bin:/opt/homebrew/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-HOME = "/tmp/fabro-trycmd-nonexistent"
-ANTHROPIC_API_KEY = "sk-test-dummy"
diff --git a/lib/crates/fabro-cli/tests/cmd/doctor/help.trycmd b/lib/crates/fabro-cli/tests/cmd/doctor/help.trycmd
deleted file mode 100644
index c6f233dac..000000000
--- a/lib/crates/fabro-cli/tests/cmd/doctor/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro doctor --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/exec/invalid-permissions.stdout b/lib/crates/fabro-cli/tests/cmd/exec/invalid-permissions.stdout
deleted file mode 100644
index 3e840343f..000000000
--- a/lib/crates/fabro-cli/tests/cmd/exec/invalid-permissions.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]invalid value[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/exec/invalid-permissions.toml b/lib/crates/fabro-cli/tests/cmd/exec/invalid-permissions.toml
deleted file mode 100644
index b5624bffc..000000000
--- a/lib/crates/fabro-cli/tests/cmd/exec/invalid-permissions.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-bin.name = "fabro"
-args = ["exec", "--permissions", "bogus", "test prompt"]
-status = "failed"
-stderr-to-stdout = true
-
-[env]
-inherit = false
diff --git a/lib/crates/fabro-cli/tests/cmd/exec/no-prompt.stdout b/lib/crates/fabro-cli/tests/cmd/exec/no-prompt.stdout
deleted file mode 100644
index 3a0e30ac7..000000000
--- a/lib/crates/fabro-cli/tests/cmd/exec/no-prompt.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]Usage:[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/exec/no-prompt.toml b/lib/crates/fabro-cli/tests/cmd/exec/no-prompt.toml
deleted file mode 100644
index 025a7d013..000000000
--- a/lib/crates/fabro-cli/tests/cmd/exec/no-prompt.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-bin.name = "fabro"
-args = ["exec"]
-status = "failed"
-stderr-to-stdout = true
-
-[env]
-inherit = false
diff --git a/lib/crates/fabro-cli/tests/cmd/init/help.trycmd b/lib/crates/fabro-cli/tests/cmd/init/help.trycmd
deleted file mode 100644
index b33b5f0c2..000000000
--- a/lib/crates/fabro-cli/tests/cmd/init/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro init --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/install/help.trycmd b/lib/crates/fabro-cli/tests/cmd/install/help.trycmd
deleted file mode 100644
index 343b5089a..000000000
--- a/lib/crates/fabro-cli/tests/cmd/install/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro install --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/llm/prompt-bad-option.stdout b/lib/crates/fabro-cli/tests/cmd/llm/prompt-bad-option.stdout
deleted file mode 100644
index 2332705a9..000000000
--- a/lib/crates/fabro-cli/tests/cmd/llm/prompt-bad-option.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]expected key=value[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/llm/prompt-bad-option.toml b/lib/crates/fabro-cli/tests/cmd/llm/prompt-bad-option.toml
deleted file mode 100644
index a2b0468bc..000000000
--- a/lib/crates/fabro-cli/tests/cmd/llm/prompt-bad-option.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-bin.name = "fabro"
-args = ["llm", "prompt", "-o", "bad_option", "hello"]
-status = "failed"
-stderr-to-stdout = true
-
-[env]
-inherit = false
diff --git a/lib/crates/fabro-cli/tests/cmd/llm/prompt-no-text.stdout b/lib/crates/fabro-cli/tests/cmd/llm/prompt-no-text.stdout
deleted file mode 100644
index 1d47217d0..000000000
--- a/lib/crates/fabro-cli/tests/cmd/llm/prompt-no-text.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]no prompt provided[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/llm/prompt-no-text.toml b/lib/crates/fabro-cli/tests/cmd/llm/prompt-no-text.toml
deleted file mode 100644
index 615ad67fc..000000000
--- a/lib/crates/fabro-cli/tests/cmd/llm/prompt-no-text.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-bin.name = "fabro"
-args = ["llm", "prompt"]
-status = "failed"
-stderr-to-stdout = true
-
-[env]
-inherit = false
diff --git a/lib/crates/fabro-cli/tests/cmd/llm/prompt-schema-invalid.stdout b/lib/crates/fabro-cli/tests/cmd/llm/prompt-schema-invalid.stdout
deleted file mode 100644
index 3aaa85b0d..000000000
--- a/lib/crates/fabro-cli/tests/cmd/llm/prompt-schema-invalid.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]--schema must be valid JSON[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/llm/prompt-schema-invalid.toml b/lib/crates/fabro-cli/tests/cmd/llm/prompt-schema-invalid.toml
deleted file mode 100644
index 5c3807e20..000000000
--- a/lib/crates/fabro-cli/tests/cmd/llm/prompt-schema-invalid.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-bin.name = "fabro"
-args = ["llm", "prompt", "--no-stream", "-m", "test-model", "--schema", "not json", "hello"]
-status = "failed"
-stderr-to-stdout = true
-
-[env]
-inherit = false
diff --git a/lib/crates/fabro-cli/tests/cmd/model/bare.trycmd b/lib/crates/fabro-cli/tests/cmd/model/bare.trycmd
deleted file mode 100644
index 9498915fb..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/bare.trycmd
+++ /dev/null
@@ -1,25 +0,0 @@
-```console
-$ fabro model
- MODEL                               PROVIDER   ALIASES                  CONTEXT            COST       SPEED 
- claude-opus-4-6                     anthropic  opus, claude-opus             1m   $15.0 / $75.0    25 tok/s 
- claude-sonnet-4-5                   anthropic                              200k    $3.0 / $15.0    50 tok/s 
- claude-sonnet-4-6                   anthropic  sonnet, claude-sonnet       200k    $3.0 / $15.0    50 tok/s 
- claude-haiku-4-5                    anthropic  haiku, claude-haiku         200k     $0.8 / $4.0   100 tok/s 
- gpt-5.2                             openai     gpt5                          1m    $1.8 / $14.0    65 tok/s 
- gpt-5-mini                          openai     gpt5-mini                     1m     $0.2 / $2.0    70 tok/s 
- gpt-5.2-codex                       openai                                   1m    $1.8 / $14.0   100 tok/s 
- gpt-5.3-codex                       openai     codex                         1m    $1.8 / $14.0   100 tok/s 
- gpt-5.3-codex-spark                 openai     codex-spark                 131k           - / -  1000 tok/s 
- gpt-5.4                             openai     gpt54, gpt-54                 1m    $2.5 / $15.0    70 tok/s 
- gpt-5.4-pro                         openai     gpt54-pro, gpt-54-pro         1m  $30.0 / $180.0    20 tok/s 
- gpt-5.4-mini                        openai     gpt54-mini, gpt-54-mini     400k     $0.8 / $4.5   140 tok/s 
- gemini-3.1-pro-preview              gemini     gemini-pro                    1m    $2.0 / $12.0    85 tok/s 
- gemini-3.1-pro-preview-customtools  gemini     gemini-customtools            1m    $2.0 / $12.0    85 tok/s 
- gemini-3-flash-preview              gemini     gemini-flash                  1m     $0.5 / $3.0   150 tok/s 
- gemini-3.1-flash-lite-preview       gemini     gemini-flash-lite             1m     $0.2 / $1.5   200 tok/s 
- kimi-k2.5                           kimi       kimi                        262k     $0.6 / $3.0    50 tok/s 
- glm-4.7                             zai        glm, glm4                   203k     $0.6 / $2.2   100 tok/s 
- minimax-m2.5                        minimax    minimax                     197k     $0.3 / $1.2    45 tok/s 
- mercury-2                           inception  mercury                     131k     $0.2 / $0.8  1000 tok/s 
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/model/help.trycmd b/lib/crates/fabro-cli/tests/cmd/model/help.trycmd
deleted file mode 100644
index b376d6c1c..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro model --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/model/list-provider.trycmd b/lib/crates/fabro-cli/tests/cmd/model/list-provider.trycmd
deleted file mode 100644
index 83d18371b..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/list-provider.trycmd
+++ /dev/null
@@ -1,9 +0,0 @@
-```console
-$ fabro model list --provider anthropic
- MODEL              PROVIDER   ALIASES                CONTEXT           COST      SPEED 
- claude-opus-4-6    anthropic  opus, claude-opus           1m  $15.0 / $75.0   25 tok/s 
- claude-sonnet-4-5  anthropic                            200k   $3.0 / $15.0   50 tok/s 
- claude-sonnet-4-6  anthropic  sonnet, claude-sonnet     200k   $3.0 / $15.0   50 tok/s 
- claude-haiku-4-5   anthropic  haiku, claude-haiku       200k    $0.8 / $4.0  100 tok/s 
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/model/list-query-aliases.trycmd b/lib/crates/fabro-cli/tests/cmd/model/list-query-aliases.trycmd
deleted file mode 100644
index 6a6ca5c05..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/list-query-aliases.trycmd
+++ /dev/null
@@ -1,8 +0,0 @@
-```console
-$ fabro model list --query codex
- MODEL                PROVIDER  ALIASES      CONTEXT          COST       SPEED 
- gpt-5.2-codex        openai                      1m  $1.8 / $14.0   100 tok/s 
- gpt-5.3-codex        openai    codex             1m  $1.8 / $14.0   100 tok/s 
- gpt-5.3-codex-spark  openai    codex-spark     131k         - / -  1000 tok/s 
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/model/list-query-case-insensitive.trycmd b/lib/crates/fabro-cli/tests/cmd/model/list-query-case-insensitive.trycmd
deleted file mode 100644
index b7b4381ee..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/list-query-case-insensitive.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro model list --query OPUS
- MODEL            PROVIDER   ALIASES            CONTEXT           COST     SPEED 
- claude-opus-4-6  anthropic  opus, claude-opus       1m  $15.0 / $75.0  25 tok/s 
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/model/list-query.trycmd b/lib/crates/fabro-cli/tests/cmd/model/list-query.trycmd
deleted file mode 100644
index e3ae96a60..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/list-query.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro model list --query opus
- MODEL            PROVIDER   ALIASES            CONTEXT           COST     SPEED 
- claude-opus-4-6  anthropic  opus, claude-opus       1m  $15.0 / $75.0  25 tok/s 
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/model/list.trycmd b/lib/crates/fabro-cli/tests/cmd/model/list.trycmd
deleted file mode 100644
index cd4c85681..000000000
--- a/lib/crates/fabro-cli/tests/cmd/model/list.trycmd
+++ /dev/null
@@ -1,25 +0,0 @@
-```console
-$ fabro model list
- MODEL                               PROVIDER   ALIASES                  CONTEXT            COST       SPEED 
- claude-opus-4-6                     anthropic  opus, claude-opus             1m   $15.0 / $75.0    25 tok/s 
- claude-sonnet-4-5                   anthropic                              200k    $3.0 / $15.0    50 tok/s 
- claude-sonnet-4-6                   anthropic  sonnet, claude-sonnet       200k    $3.0 / $15.0    50 tok/s 
- claude-haiku-4-5                    anthropic  haiku, claude-haiku         200k     $0.8 / $4.0   100 tok/s 
- gpt-5.2                             openai     gpt5                          1m    $1.8 / $14.0    65 tok/s 
- gpt-5-mini                          openai     gpt5-mini                     1m     $0.2 / $2.0    70 tok/s 
- gpt-5.2-codex                       openai                                   1m    $1.8 / $14.0   100 tok/s 
- gpt-5.3-codex                       openai     codex                         1m    $1.8 / $14.0   100 tok/s 
- gpt-5.3-codex-spark                 openai     codex-spark                 131k           - / -  1000 tok/s 
- gpt-5.4                             openai     gpt54, gpt-54                 1m    $2.5 / $15.0    70 tok/s 
- gpt-5.4-pro                         openai     gpt54-pro, gpt-54-pro         1m  $30.0 / $180.0    20 tok/s 
- gpt-5.4-mini                        openai     gpt54-mini, gpt-54-mini     400k     $0.8 / $4.5   140 tok/s 
- gemini-3.1-pro-preview              gemini     gemini-pro                    1m    $2.0 / $12.0    85 tok/s 
- gemini-3.1-pro-preview-customtools  gemini     gemini-customtools            1m    $2.0 / $12.0    85 tok/s 
- gemini-3-flash-preview              gemini     gemini-flash                  1m     $0.5 / $3.0   150 tok/s 
- gemini-3.1-flash-lite-preview       gemini     gemini-flash-lite             1m     $0.2 / $1.5   200 tok/s 
- kimi-k2.5                           kimi       kimi                        262k     $0.6 / $3.0    50 tok/s 
- glm-4.7                             zai        glm, glm4                   203k     $0.6 / $2.2   100 tok/s 
- minimax-m2.5                        minimax    minimax                     197k     $0.3 / $1.2    45 tok/s 
- mercury-2                           inception  mercury                     131k     $0.2 / $0.8  1000 tok/s 
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/pr/help.trycmd b/lib/crates/fabro-cli/tests/cmd/pr/help.trycmd
deleted file mode 100644
index 3afd70ab3..000000000
--- a/lib/crates/fabro-cli/tests/cmd/pr/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro pr --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/preview/help.trycmd b/lib/crates/fabro-cli/tests/cmd/preview/help.trycmd
deleted file mode 100644
index 6ff861312..000000000
--- a/lib/crates/fabro-cli/tests/cmd/preview/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro preview --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/repo/help.trycmd b/lib/crates/fabro-cli/tests/cmd/repo/help.trycmd
deleted file mode 100644
index 5fe17b3da..000000000
--- a/lib/crates/fabro-cli/tests/cmd/repo/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro repo --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/run/dry-run-branching.toml b/lib/crates/fabro-cli/tests/cmd/run/dry-run-branching.toml
deleted file mode 100644
index dc18a4f72..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/dry-run-branching.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["run", "--dry-run", "--auto-approve", "../../../test/branching.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/run/dry-run-conditions.toml b/lib/crates/fabro-cli/tests/cmd/run/dry-run-conditions.toml
deleted file mode 100644
index be6796457..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/dry-run-conditions.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["run", "--dry-run", "--auto-approve", "../../../test/conditions.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/run/dry-run-legacy-tool.toml b/lib/crates/fabro-cli/tests/cmd/run/dry-run-legacy-tool.toml
deleted file mode 100644
index d8f0822ec..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/dry-run-legacy-tool.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["run", "--dry-run", "--auto-approve", "../../../test/legacy_tool.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/run/dry-run-parallel.toml b/lib/crates/fabro-cli/tests/cmd/run/dry-run-parallel.toml
deleted file mode 100644
index bb98ed609..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/dry-run-parallel.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["run", "--dry-run", "--auto-approve", "../../../test/parallel.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/run/dry-run-simple.toml b/lib/crates/fabro-cli/tests/cmd/run/dry-run-simple.toml
deleted file mode 100644
index d304c33da..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/dry-run-simple.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["run", "--dry-run", "--auto-approve", "../../../test/simple.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/run/dry-run-styled.toml b/lib/crates/fabro-cli/tests/cmd/run/dry-run-styled.toml
deleted file mode 100644
index c40ec2860..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/dry-run-styled.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["run", "--dry-run", "--auto-approve", "../../../test/styled.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/run/help.trycmd b/lib/crates/fabro-cli/tests/cmd/run/help.trycmd
deleted file mode 100644
index 5f48728ef..000000000
--- a/lib/crates/fabro-cli/tests/cmd/run/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro run --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/ssh/help.trycmd b/lib/crates/fabro-cli/tests/cmd/ssh/help.trycmd
deleted file mode 100644
index 2a7ad36c0..000000000
--- a/lib/crates/fabro-cli/tests/cmd/ssh/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro ssh --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/system/help.trycmd b/lib/crates/fabro-cli/tests/cmd/system/help.trycmd
deleted file mode 100644
index 7415294d4..000000000
--- a/lib/crates/fabro-cli/tests/cmd/system/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro system --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/top-level/no-dotenv-flag.stdout b/lib/crates/fabro-cli/tests/cmd/top-level/no-dotenv-flag.stdout
deleted file mode 100644
index 42412c762..000000000
--- a/lib/crates/fabro-cli/tests/cmd/top-level/no-dotenv-flag.stdout
+++ /dev/null
@@ -1,3 +0,0 @@
-...
-[..]unexpected argument[..]
-...
diff --git a/lib/crates/fabro-cli/tests/cmd/top-level/no-dotenv-flag.toml b/lib/crates/fabro-cli/tests/cmd/top-level/no-dotenv-flag.toml
deleted file mode 100644
index b852d9b3f..000000000
--- a/lib/crates/fabro-cli/tests/cmd/top-level/no-dotenv-flag.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-bin.name = "fabro"
-args = ["--no-dotenv", "doctor"]
-status = "failed"
-stderr-to-stdout = true
-
-[env]
-inherit = false
diff --git a/lib/crates/fabro-cli/tests/cmd/top-level/version.trycmd b/lib/crates/fabro-cli/tests/cmd/top-level/version.trycmd
deleted file mode 100644
index 3de33a6f6..000000000
--- a/lib/crates/fabro-cli/tests/cmd/top-level/version.trycmd
+++ /dev/null
@@ -1,5 +0,0 @@
-```console
-$ fabro --version
-fabro [..]
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/branching.toml b/lib/crates/fabro-cli/tests/cmd/validate/branching.toml
deleted file mode 100644
index 90ed3679c..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/branching.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/branching.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/conditions.toml b/lib/crates/fabro-cli/tests/cmd/validate/conditions.toml
deleted file mode 100644
index ca87ea280..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/conditions.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/conditions.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/help.trycmd b/lib/crates/fabro-cli/tests/cmd/validate/help.trycmd
deleted file mode 100644
index ca53c3141..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/help.trycmd
+++ /dev/null
@@ -1,6 +0,0 @@
-```console
-$ fabro validate --help
-[..]
-...
-
-```
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/invalid.toml b/lib/crates/fabro-cli/tests/cmd/validate/invalid.toml
deleted file mode 100644
index f2f9bce1f..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/invalid.toml
+++ /dev/null
@@ -1,10 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/invalid.fabro"]
-status = "failed"
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/legacy-tool.toml b/lib/crates/fabro-cli/tests/cmd/validate/legacy-tool.toml
deleted file mode 100644
index 9b63836f2..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/legacy-tool.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/legacy_tool.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/parallel.toml b/lib/crates/fabro-cli/tests/cmd/validate/parallel.toml
deleted file mode 100644
index e27985d3e..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/parallel.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/parallel.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/simple.toml b/lib/crates/fabro-cli/tests/cmd/validate/simple.toml
deleted file mode 100644
index de446a3b7..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/simple.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/simple.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/cmd/validate/styled.toml b/lib/crates/fabro-cli/tests/cmd/validate/styled.toml
deleted file mode 100644
index 745bf7d6f..000000000
--- a/lib/crates/fabro-cli/tests/cmd/validate/styled.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-bin.name = "fabro"
-args = ["validate", "../../../test/styled.fabro"]
-
-[env]
-inherit = false
-
-[fs]
-sandbox = false
-cwd = "../../.."
diff --git a/lib/crates/fabro-cli/tests/it/cmd/config.rs b/lib/crates/fabro-cli/tests/it/cmd/config.rs
new file mode 100644
index 000000000..fa16d1c3a
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/config.rs
@@ -0,0 +1,56 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.config();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Inspect merged configuration
+
+    Usage: fabro config [OPTIONS] <COMMAND>
+
+    Commands:
+      show  Print the merged FabroSettings as YAML
+      help  Print this message or the help of the given subcommand(s)
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn show_help() {
+    let context = test_context!();
+    let mut cmd = context.config();
+    cmd.args(["show", "--help"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Print the merged FabroSettings as YAML
+
+    Usage: fabro config show [OPTIONS] [WORKFLOW]
+
+    Arguments:
+      [WORKFLOW]  Optional workflow name, .fabro path, or .toml run config to overlay
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/cp.rs b/lib/crates/fabro-cli/tests/it/cmd/cp.rs
new file mode 100644
index 000000000..7f2a1f2b6
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/cp.rs
@@ -0,0 +1,30 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.cp();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Copy files to/from a run's sandbox
+
+    Usage: fabro cp [OPTIONS] <SRC> <DST>
+
+    Arguments:
+      <SRC>  Source: <run-id>:<path> or local path
+      <DST>  Destination: <run-id>:<path> or local path
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+      -r, --recursive                  Recurse into directories
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/doctor.rs b/lib/crates/fabro-cli/tests/it/cmd/doctor.rs
new file mode 100644
index 000000000..aa843e0b7
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/doctor.rs
@@ -0,0 +1,62 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.doctor();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Check environment and integration health
+
+    Usage: fabro doctor [OPTIONS]
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+      -v, --verbose                    Show detailed information for each check
+          --dry-run                    Skip live service probes (LLM, sandbox, API, web, Brave Search)
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn dry_run_flag() {
+    let context = test_context!();
+    let mut cmd = context.doctor();
+    cmd.arg("--dry-run");
+    cmd.env(
+        "PATH",
+        "/usr/local/bin:/opt/homebrew/bin:/usr/bin:/bin:/usr/sbin:/sbin",
+    );
+    cmd.env("ANTHROPIC_API_KEY", "sk-test-dummy");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Fabro Doctor
+
+      Required
+      [!] Configuration (no user config file found)
+      [✓] LLM providers (1 configured)
+      [!] GitHub App (not configured)
+
+      Optional
+      [!] Cloud sandbox (no sandbox configured)
+      [!] Brave Search (not configured)
+
+    Found issues in 4 categories.
+
+    Warnings:
+      • Configuration — Create ~/.fabro/user.toml
+      • GitHub App — Configure GitHub App in server.toml and set env vars to enable GitHub integration
+      • Cloud sandbox — Set DAYTONA_API_KEY to enable cloud sandbox execution
+      • Brave Search — Set BRAVE_SEARCH_API_KEY to enable web search
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/exec.rs b/lib/crates/fabro-cli/tests/it/cmd/exec.rs
new file mode 100644
index 000000000..d9dcb9243
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/exec.rs
@@ -0,0 +1,35 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn invalid_permissions() {
+    let context = test_context!();
+    let mut cmd = context.exec_cmd();
+    cmd.args(["--permissions", "bogus", "test prompt"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: false
+    exit_code: 2
+    ----- stdout -----
+    ----- stderr -----
+    error: invalid value 'bogus' for '--permissions <PERMISSIONS>'
+      [possible values: read-only, read-write, full]
+
+    For more information, try '--help'.
+    ");
+}
+
+#[test]
+fn no_prompt() {
+    let context = test_context!();
+    fabro_snapshot!(context.filters(), context.exec_cmd(), @"
+    success: false
+    exit_code: 2
+    ----- stdout -----
+    ----- stderr -----
+    error: the following required arguments were not provided:
+      <PROMPT>
+
+    Usage: fabro exec --no-upgrade-check --storage-dir <STORAGE_DIR> <PROMPT>
+
+    For more information, try '--help'.
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/init.rs b/lib/crates/fabro-cli/tests/it/cmd/init.rs
new file mode 100644
index 000000000..560047540
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/init.rs
@@ -0,0 +1,25 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.init_cmd();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Initialize a new project (deprecated: use `repo init`)
+
+    Usage: fabro init [OPTIONS]
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/install.rs b/lib/crates/fabro-cli/tests/it/cmd/install.rs
new file mode 100644
index 000000000..04988166a
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/install.rs
@@ -0,0 +1,26 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.install();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Set up the Fabro environment (LLMs, certs, GitHub)
+
+    Usage: fabro install [OPTIONS]
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --web-url <WEB_URL>          Base URL for the web UI (used for OAuth callback URLs) [default: http://localhost:5173]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/llm.rs b/lib/crates/fabro-cli/tests/it/cmd/llm.rs
new file mode 100644
index 000000000..cb2d3ccf2
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/llm.rs
@@ -0,0 +1,55 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn prompt_bad_option() {
+    let context = test_context!();
+    let mut cmd = context.llm();
+    cmd.args(["prompt", "-o", "bad_option", "hello"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: false
+    exit_code: 2
+    ----- stdout -----
+    ----- stderr -----
+    error: invalid value 'bad_option' for '--option <OPTION>': expected key=value, got bad_option
+
+    For more information, try '--help'.
+    ");
+}
+
+#[test]
+fn prompt_no_text() {
+    let context = test_context!();
+    let mut cmd = context.llm();
+    cmd.arg("prompt");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: false
+    exit_code: 1
+    ----- stdout -----
+    ----- stderr -----
+    error: Error: no prompt provided. Pass a prompt as an argument or pipe text via stdin.
+    ");
+}
+
+#[test]
+fn prompt_schema_invalid() {
+    let context = test_context!();
+    let mut cmd = context.llm();
+    cmd.args([
+        "prompt",
+        "--no-stream",
+        "-m",
+        "test-model",
+        "--schema",
+        "not json",
+        "hello",
+    ]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: false
+    exit_code: 1
+    ----- stdout -----
+    ----- stderr -----
+    Using model: test-model
+    error: --schema must be valid JSON
+      > expected ident at line 1 column 2
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/mod.rs b/lib/crates/fabro-cli/tests/it/cmd/mod.rs
new file mode 100644
index 000000000..0f3b59762
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/mod.rs
@@ -0,0 +1,17 @@
+mod config;
+mod cp;
+mod doctor;
+mod exec;
+mod init;
+mod install;
+mod llm;
+mod model;
+mod pr;
+mod preview;
+mod repo;
+mod run;
+mod secret;
+mod ssh;
+mod system;
+mod top_level;
+mod validate;
diff --git a/lib/crates/fabro-cli/tests/it/cmd/model.rs b/lib/crates/fabro-cli/tests/it/cmd/model.rs
new file mode 100644
index 000000000..98ae48c4f
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/model.rs
@@ -0,0 +1,161 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.model();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    List and test LLM models
+
+    Usage: fabro model [OPTIONS] [COMMAND]
+
+    Commands:
+      list  List available models
+      test  Test model availability by sending a simple prompt
+      help  Print this message or the help of the given subcommand(s)
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn bare() {
+    let context = test_context!();
+    fabro_snapshot!(context.filters(), context.model(), @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+     MODEL                               PROVIDER   ALIASES                  CONTEXT            COST       SPEED 
+     claude-opus-4-6                     anthropic  opus, claude-opus             1m   $15.0 / $75.0    25 tok/s 
+     claude-sonnet-4-5                   anthropic                              200k    $3.0 / $15.0    50 tok/s 
+     claude-sonnet-4-6                   anthropic  sonnet, claude-sonnet       200k    $3.0 / $15.0    50 tok/s 
+     claude-haiku-4-5                    anthropic  haiku, claude-haiku         200k     $0.8 / $4.0   100 tok/s 
+     gpt-5.2                             openai     gpt5                          1m    $1.8 / $14.0    65 tok/s 
+     gpt-5-mini                          openai     gpt5-mini                     1m     $0.2 / $2.0    70 tok/s 
+     gpt-5.2-codex                       openai                                   1m    $1.8 / $14.0   100 tok/s 
+     gpt-5.3-codex                       openai     codex                         1m    $1.8 / $14.0   100 tok/s 
+     gpt-5.3-codex-spark                 openai     codex-spark                 131k           - / -  1000 tok/s 
+     gpt-5.4                             openai     gpt54, gpt-54                 1m    $2.5 / $15.0    70 tok/s 
+     gpt-5.4-pro                         openai     gpt54-pro, gpt-54-pro         1m  $30.0 / $180.0    20 tok/s 
+     gpt-5.4-mini                        openai     gpt54-mini, gpt-54-mini     400k     $0.8 / $4.5   140 tok/s 
+     gemini-3.1-pro-preview              gemini     gemini-pro                    1m    $2.0 / $12.0    85 tok/s 
+     gemini-3.1-pro-preview-customtools  gemini     gemini-customtools            1m    $2.0 / $12.0    85 tok/s 
+     gemini-3-flash-preview              gemini     gemini-flash                  1m     $0.5 / $3.0   150 tok/s 
+     gemini-3.1-flash-lite-preview       gemini     gemini-flash-lite             1m     $0.2 / $1.5   200 tok/s 
+     kimi-k2.5                           kimi       kimi                        262k     $0.6 / $3.0    50 tok/s 
+     glm-4.7                             zai        glm, glm4                   203k     $0.6 / $2.2   100 tok/s 
+     minimax-m2.5                        minimax    minimax                     197k     $0.3 / $1.2    45 tok/s 
+     mercury-2                           inception  mercury                     131k     $0.2 / $0.8  1000 tok/s 
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn list() {
+    let context = test_context!();
+    let mut cmd = context.model();
+    cmd.arg("list");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+     MODEL                               PROVIDER   ALIASES                  CONTEXT            COST       SPEED 
+     claude-opus-4-6                     anthropic  opus, claude-opus             1m   $15.0 / $75.0    25 tok/s 
+     claude-sonnet-4-5                   anthropic                              200k    $3.0 / $15.0    50 tok/s 
+     claude-sonnet-4-6                   anthropic  sonnet, claude-sonnet       200k    $3.0 / $15.0    50 tok/s 
+     claude-haiku-4-5                    anthropic  haiku, claude-haiku         200k     $0.8 / $4.0   100 tok/s 
+     gpt-5.2                             openai     gpt5                          1m    $1.8 / $14.0    65 tok/s 
+     gpt-5-mini                          openai     gpt5-mini                     1m     $0.2 / $2.0    70 tok/s 
+     gpt-5.2-codex                       openai                                   1m    $1.8 / $14.0   100 tok/s 
+     gpt-5.3-codex                       openai     codex                         1m    $1.8 / $14.0   100 tok/s 
+     gpt-5.3-codex-spark                 openai     codex-spark                 131k           - / -  1000 tok/s 
+     gpt-5.4                             openai     gpt54, gpt-54                 1m    $2.5 / $15.0    70 tok/s 
+     gpt-5.4-pro                         openai     gpt54-pro, gpt-54-pro         1m  $30.0 / $180.0    20 tok/s 
+     gpt-5.4-mini                        openai     gpt54-mini, gpt-54-mini     400k     $0.8 / $4.5   140 tok/s 
+     gemini-3.1-pro-preview              gemini     gemini-pro                    1m    $2.0 / $12.0    85 tok/s 
+     gemini-3.1-pro-preview-customtools  gemini     gemini-customtools            1m    $2.0 / $12.0    85 tok/s 
+     gemini-3-flash-preview              gemini     gemini-flash                  1m     $0.5 / $3.0   150 tok/s 
+     gemini-3.1-flash-lite-preview       gemini     gemini-flash-lite             1m     $0.2 / $1.5   200 tok/s 
+     kimi-k2.5                           kimi       kimi                        262k     $0.6 / $3.0    50 tok/s 
+     glm-4.7                             zai        glm, glm4                   203k     $0.6 / $2.2   100 tok/s 
+     minimax-m2.5                        minimax    minimax                     197k     $0.3 / $1.2    45 tok/s 
+     mercury-2                           inception  mercury                     131k     $0.2 / $0.8  1000 tok/s 
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn list_provider() {
+    let context = test_context!();
+    let mut cmd = context.model();
+    cmd.args(["list", "--provider", "anthropic"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+     MODEL              PROVIDER   ALIASES                CONTEXT           COST      SPEED 
+     claude-opus-4-6    anthropic  opus, claude-opus           1m  $15.0 / $75.0   25 tok/s 
+     claude-sonnet-4-5  anthropic                            200k   $3.0 / $15.0   50 tok/s 
+     claude-sonnet-4-6  anthropic  sonnet, claude-sonnet     200k   $3.0 / $15.0   50 tok/s 
+     claude-haiku-4-5   anthropic  haiku, claude-haiku       200k    $0.8 / $4.0  100 tok/s 
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn list_query() {
+    let context = test_context!();
+    let mut cmd = context.model();
+    cmd.args(["list", "--query", "opus"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+     MODEL            PROVIDER   ALIASES            CONTEXT           COST     SPEED 
+     claude-opus-4-6  anthropic  opus, claude-opus       1m  $15.0 / $75.0  25 tok/s 
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn list_query_aliases() {
+    let context = test_context!();
+    let mut cmd = context.model();
+    cmd.args(["list", "--query", "codex"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+     MODEL                PROVIDER  ALIASES      CONTEXT          COST       SPEED 
+     gpt-5.2-codex        openai                      1m  $1.8 / $14.0   100 tok/s 
+     gpt-5.3-codex        openai    codex             1m  $1.8 / $14.0   100 tok/s 
+     gpt-5.3-codex-spark  openai    codex-spark     131k         - / -  1000 tok/s 
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn list_query_case_insensitive() {
+    let context = test_context!();
+    let mut cmd = context.model();
+    cmd.args(["list", "--query", "OPUS"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+     MODEL            PROVIDER   ALIASES            CONTEXT           COST     SPEED 
+     claude-opus-4-6  anthropic  opus, claude-opus       1m  $15.0 / $75.0  25 tok/s 
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/pr.rs b/lib/crates/fabro-cli/tests/it/cmd/pr.rs
new file mode 100644
index 000000000..441fc66bc
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/pr.rs
@@ -0,0 +1,33 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.pr();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Pull request operations
+
+    Usage: fabro pr [OPTIONS] <COMMAND>
+
+    Commands:
+      create  Create a pull request from a completed run
+      list    List pull requests from workflow runs
+      view    View pull request details
+      merge   Merge a pull request
+      close   Close a pull request
+      help    Print this message or the help of the given subcommand(s)
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/preview.rs b/lib/crates/fabro-cli/tests/it/cmd/preview.rs
new file mode 100644
index 000000000..7b11aa0ff
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/preview.rs
@@ -0,0 +1,32 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.preview();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Get a preview URL for a port on a run's sandbox
+
+    Usage: fabro preview [OPTIONS] <RUN> <PORT>
+
+    Arguments:
+      <RUN>   Run ID or prefix
+      <PORT>  Port number
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --signed                     Generate a signed URL (embeds auth token, no headers needed)
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --ttl <TTL>                  Signed URL expiry in seconds (default 3600, requires --signed) [default: 3600]
+          --open                       Open URL in browser (implies --signed)
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/repo.rs b/lib/crates/fabro-cli/tests/it/cmd/repo.rs
new file mode 100644
index 000000000..6ff144c87
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/repo.rs
@@ -0,0 +1,30 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.repo();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Repository commands
+
+    Usage: fabro repo [OPTIONS] <COMMAND>
+
+    Commands:
+      init    Initialize a new project
+      deinit  Remove fabro.toml and fabro/ directory
+      help    Print this message or the help of the given subcommand(s)
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/run.rs b/lib/crates/fabro-cli/tests/it/cmd/run.rs
new file mode 100644
index 000000000..8bab3b2d8
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/run.rs
@@ -0,0 +1,236 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+fn fixture(name: &str) -> std::path::PathBuf {
+    std::path::Path::new(env!("CARGO_MANIFEST_DIR")).join(format!("../../../test/{name}"))
+}
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Launch a workflow run
+
+    Usage: fabro run [OPTIONS] <WORKFLOW>
+
+    Arguments:
+      <WORKFLOW>  Path to a .fabro workflow file or .toml task config
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --dry-run                    Execute with simulated LLM backend
+          --auto-approve               Auto-approve all human gates
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --goal <GOAL>                Override the workflow goal (exposed as $goal in prompts)
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --goal-file <GOAL_FILE>      Read the workflow goal from a file
+          --model <MODEL>              Override default LLM model
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+          --provider <PROVIDER>        Override default LLM provider
+      -v, --verbose                    Enable verbose output
+          --sandbox <SANDBOX>          Sandbox for agent tools [possible values: local, docker, daytona, ssh]
+          --label <KEY=VALUE>          Attach a label to this run (repeatable, format: KEY=VALUE)
+          --no-retro                   Skip retro generation after the run
+          --preserve-sandbox           Keep the sandbox alive after the run finishes (for debugging)
+      -d, --detach                     Run the workflow in the background and print the run ID
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn dry_run_simple() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.args(["--dry-run", "--auto-approve"]);
+    cmd.arg(fixture("simple.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Simple (4 nodes, 3 edges)
+    Graph: ../../../test/simple.fabro
+    Goal: Run tests and report results
+
+        Sandbox: local (ready in [TIME])
+        ✓ Start  0ms
+        ✓ Run Tests  0ms
+        ✓ Report  0ms
+        ✓ Exit  0ms
+
+    === Run Result ===
+    Run:       [ULID]
+    Status:    SUCCESS
+    Duration:  0 seconds
+    Run:       [STORAGE_DIR]/runs/20260330-dry-run-[ULID]
+
+    === Output ===
+    [Simulated] Response for stage: report
+    ");
+}
+
+#[test]
+fn dry_run_branching() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.args(["--dry-run", "--auto-approve"]);
+    cmd.arg(fixture("branching.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Branch (6 nodes, 6 edges)
+    Graph: ../../../test/branching.fabro
+    Goal: Implement and validate a feature
+
+    warning [node: implement]: Node 'implement' has goal_gate=true but no retry_target or fallback_retry_target (goal_gate_has_retry)
+        Sandbox: local (ready in [TIME])
+        ✓ Start  0ms
+        ✓ Plan  0ms
+        ✓ Implement  0ms
+        ✓ Validate  0ms
+        ✓ Tests passing?  0ms
+        ✓ Exit  0ms
+
+    === Run Result ===
+    Run:       [ULID]
+    Status:    SUCCESS
+    Duration:  0 seconds
+    Run:       [STORAGE_DIR]/runs/20260330-dry-run-[ULID]
+
+    === Output ===
+    [Simulated] Response for stage: validate
+    ");
+}
+
+#[test]
+fn dry_run_conditions() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.args(["--dry-run", "--auto-approve"]);
+    cmd.arg(fixture("conditions.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Conditions (5 nodes, 5 edges)
+    Graph: ../../../test/conditions.fabro
+    Goal: Test condition evaluation with OR and parentheses
+
+        Sandbox: local (ready in [TIME])
+        ✓ start  0ms
+        ✓ Decide  0ms
+        ✓ Path B  0ms
+        ✓ exit  0ms
+
+    === Run Result ===
+    Run:       [ULID]
+    Status:    SUCCESS
+    Duration:  0 seconds
+    Run:       [STORAGE_DIR]/runs/20260330-dry-run-[ULID]
+
+    === Output ===
+    [Simulated] Response for stage: path_b
+    ");
+}
+
+#[test]
+fn dry_run_parallel() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.args(["--dry-run", "--auto-approve"]);
+    cmd.arg(fixture("parallel.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Parallel (7 nodes, 7 edges)
+    Graph: ../../../test/parallel.fabro
+    Goal: Test parallel and fan-in execution
+
+        Sandbox: local (ready in [TIME])
+        ✓ start  0ms
+        ✓ Fork Work  0ms
+        ✓ Merge Results  0ms
+        ✓ Review  0ms
+        ✓ exit  0ms
+
+    === Run Result ===
+    Run:       [ULID]
+    Status:    SUCCESS
+    Duration:  0 seconds
+    Run:       [STORAGE_DIR]/runs/20260330-dry-run-[ULID]
+
+    === Output ===
+    [Simulated] Response for stage: review
+    ");
+}
+
+#[test]
+fn dry_run_styled() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.args(["--dry-run", "--auto-approve"]);
+    cmd.arg(fixture("styled.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Styled (5 nodes, 4 edges)
+    Graph: ../../../test/styled.fabro
+    Goal: Build a styled pipeline
+
+        Sandbox: local (ready in [TIME])
+        ✓ start  0ms
+        ✓ Plan  0ms
+        ✓ Implement  0ms
+        ✓ Critical Review  0ms
+        ✓ exit  0ms
+
+    === Run Result ===
+    Run:       [ULID]
+    Status:    SUCCESS
+    Duration:  0 seconds
+    Run:       [STORAGE_DIR]/runs/20260330-dry-run-[ULID]
+
+    === Output ===
+    [Simulated] Response for stage: critical_review
+    ");
+}
+
+#[test]
+fn dry_run_legacy_tool() {
+    let context = test_context!();
+    let mut cmd = context.run_cmd();
+    cmd.args(["--dry-run", "--auto-approve"]);
+    cmd.arg(fixture("legacy_tool.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: LegacyTool (3 nodes, 2 edges)
+    Graph: ../../../test/legacy_tool.fabro
+    Goal: Verify backwards compatibility with old tool naming
+
+        Sandbox: local (ready in [TIME])
+        ✓ Start  0ms
+        ✓ Echo  0ms
+        ✓ Exit  0ms
+
+    === Run Result ===
+    Run:       [ULID]
+    Status:    SUCCESS
+    Duration:  0 seconds
+    Run:       [STORAGE_DIR]/runs/20260330-dry-run-[ULID]
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/secret.rs b/lib/crates/fabro-cli/tests/it/cmd/secret.rs
new file mode 100644
index 000000000..171080997
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/secret.rs
@@ -0,0 +1,32 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.secret();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Manage secrets in ~/.fabro/.env
+
+    Usage: fabro secret [OPTIONS] <COMMAND>
+
+    Commands:
+      get   Get a secret value
+      list  List secret names
+      rm    Remove a secret
+      set   Set a secret value
+      help  Print this message or the help of the given subcommand(s)
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/ssh.rs b/lib/crates/fabro-cli/tests/it/cmd/ssh.rs
new file mode 100644
index 000000000..0615aa198
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/ssh.rs
@@ -0,0 +1,30 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.ssh();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    SSH into a run's Daytona sandbox
+
+    Usage: fabro ssh [OPTIONS] <RUN>
+
+    Arguments:
+      <RUN>  Run ID or prefix
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --ttl <TTL>                  SSH access expiry in minutes (default 60) [default: 60]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --print                      Print the SSH command instead of connecting
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/system.rs b/lib/crates/fabro-cli/tests/it/cmd/system.rs
new file mode 100644
index 000000000..f57dc004d
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/system.rs
@@ -0,0 +1,30 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.system();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    System maintenance commands
+
+    Usage: fabro system [OPTIONS] <COMMAND>
+
+    Commands:
+      prune  Delete old workflow runs
+      df     Show disk usage
+      help   Print this message or the help of the given subcommand(s)
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/top_level.rs b/lib/crates/fabro-cli/tests/it/cmd/top_level.rs
new file mode 100644
index 000000000..6e9fc7a4f
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/top_level.rs
@@ -0,0 +1,33 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+#[test]
+fn version() {
+    let context = test_context!();
+    let mut cmd = context.command();
+    cmd.arg("--version");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    fabro [VERSION] (c6ff6c9 2026-03-30)
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn no_dotenv_flag() {
+    let context = test_context!();
+    let mut cmd = context.command();
+    cmd.args(["--no-dotenv", "doctor"]);
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: false
+    exit_code: 2
+    ----- stdout -----
+    ----- stderr -----
+    error: unexpected argument '--no-dotenv' found
+
+    Usage: fabro [OPTIONS] <COMMAND>
+
+    For more information, try '--help'.
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/cmd/validate.rs b/lib/crates/fabro-cli/tests/it/cmd/validate.rs
new file mode 100644
index 000000000..721870110
--- /dev/null
+++ b/lib/crates/fabro-cli/tests/it/cmd/validate.rs
@@ -0,0 +1,147 @@
+use fabro_test::{fabro_snapshot, test_context};
+
+fn fixture(name: &str) -> std::path::PathBuf {
+    std::path::Path::new(env!("CARGO_MANIFEST_DIR")).join(format!("../../../test/{name}"))
+}
+
+#[test]
+fn help() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg("--help");
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    Validate a workflow
+
+    Usage: fabro validate [OPTIONS] <WORKFLOW>
+
+    Arguments:
+      <WORKFLOW>  Path to the .fabro workflow file
+
+    Options:
+          --debug                      Enable DEBUG-level logging (default is INFO) [env: FABRO_DEBUG=]
+          --no-upgrade-check           Disable automatic upgrade check [env: FABRO_NO_UPGRADE_CHECK=true]
+          --quiet                      Suppress non-essential output [env: FABRO_QUIET=]
+          --verbose                    Enable verbose output [env: FABRO_VERBOSE=]
+          --storage-dir <STORAGE_DIR>  Storage directory (default: ~/.fabro) [env: FABRO_STORAGE_DIR=[STORAGE_DIR]]
+      -h, --help                       Print help
+    ----- stderr -----
+    ");
+}
+
+#[test]
+fn simple() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("simple.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Simple (4 nodes, 3 edges)
+    Graph: ../../../test/simple.fabro
+    Validation: OK
+    ");
+}
+
+#[test]
+fn branching() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("branching.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Branch (6 nodes, 6 edges)
+    Graph: ../../../test/branching.fabro
+    warning [node: implement]: Node 'implement' has goal_gate=true but no retry_target or fallback_retry_target (goal_gate_has_retry)
+    Validation: OK
+    ");
+}
+
+#[test]
+fn conditions() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("conditions.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Conditions (5 nodes, 5 edges)
+    Graph: ../../../test/conditions.fabro
+    Validation: OK
+    ");
+}
+
+#[test]
+fn parallel() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("parallel.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Parallel (7 nodes, 7 edges)
+    Graph: ../../../test/parallel.fabro
+    Validation: OK
+    ");
+}
+
+#[test]
+fn styled() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("styled.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Styled (5 nodes, 4 edges)
+    Graph: ../../../test/styled.fabro
+    Validation: OK
+    ");
+}
+
+#[test]
+fn legacy_tool() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("legacy_tool.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: true
+    exit_code: 0
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: LegacyTool (3 nodes, 2 edges)
+    Graph: ../../../test/legacy_tool.fabro
+    Validation: OK
+    ");
+}
+
+#[test]
+fn invalid() {
+    let context = test_context!();
+    let mut cmd = context.validate();
+    cmd.arg(fixture("invalid.fabro"));
+    fabro_snapshot!(context.filters(), cmd, @"
+    success: false
+    exit_code: 1
+    ----- stdout -----
+    ----- stderr -----
+    Workflow: Invalid (2 nodes, 1 edges)
+    Graph: ../../../test/invalid.fabro
+    error: Pipeline must have exactly one start node (shape=Mdiamond or id start/Start) (start_node)
+    error [node: exit]: Exit node 'exit' has 1 outgoing edge(s) but must have none (exit_no_outgoing)
+    error: Validation failed
+    ");
+}
diff --git a/lib/crates/fabro-cli/tests/it/main.rs b/lib/crates/fabro-cli/tests/it/main.rs
index 5887bfa77..c30a35a2c 100644
--- a/lib/crates/fabro-cli/tests/it/main.rs
+++ b/lib/crates/fabro-cli/tests/it/main.rs
@@ -1,3 +1,3 @@
 mod cli;
+mod cmd;
 mod scenario;
-mod trycmd;
diff --git a/lib/crates/fabro-cli/tests/it/trycmd.rs b/lib/crates/fabro-cli/tests/it/trycmd.rs
deleted file mode 100644
index bcacf9dd1..000000000
--- a/lib/crates/fabro-cli/tests/it/trycmd.rs
+++ /dev/null
@@ -1,30 +0,0 @@
-macro_rules! trycmd_subcommand {
-    ($name:ident, $dir:expr) => {
-        #[test]
-        fn $name() {
-            let home = tempfile::tempdir().unwrap();
-            let home = home.path().display().to_string();
-            trycmd::TestCases::new()
-                .env("HOME", home)
-                .case(concat!("tests/cmd/", $dir, "/*.trycmd"))
-                .case(concat!("tests/cmd/", $dir, "/*.toml"));
-        }
-    };
-}
-
-trycmd_subcommand!(cli_cp, "cp");
-trycmd_subcommand!(cli_config, "config");
-trycmd_subcommand!(cli_doctor, "doctor");
-trycmd_subcommand!(cli_exec, "exec");
-trycmd_subcommand!(cli_init, "init");
-trycmd_subcommand!(cli_install, "install");
-trycmd_subcommand!(cli_llm, "llm");
-trycmd_subcommand!(cli_model, "model");
-trycmd_subcommand!(cli_pr, "pr");
-trycmd_subcommand!(cli_repo, "repo");
-trycmd_subcommand!(cli_preview, "preview");
-trycmd_subcommand!(cli_run, "run");
-trycmd_subcommand!(cli_ssh, "ssh");
-trycmd_subcommand!(cli_system, "system");
-trycmd_subcommand!(cli_top_level, "top-level");
-trycmd_subcommand!(cli_validate, "validate");
diff --git a/lib/crates/fabro-test/Cargo.toml b/lib/crates/fabro-test/Cargo.toml
new file mode 100644
index 000000000..b86048174
--- /dev/null
+++ b/lib/crates/fabro-test/Cargo.toml
@@ -0,0 +1,19 @@
+[package]
+name = "fabro-test"
+edition.workspace = true
+version.workspace = true
+publish = false
+license.workspace = true
+description = "Test utilities for fabro integration tests"
+
+[lib]
+doctest = false
+
+[lints]
+workspace = true
+
+[dependencies]
+assert_cmd = "2"
+insta = { workspace = true, features = ["filters"] }
+regex = { workspace = true }
+tempfile = "3"
diff --git a/lib/crates/fabro-test/src/lib.rs b/lib/crates/fabro-test/src/lib.rs
new file mode 100644
index 000000000..ae17cd8db
--- /dev/null
+++ b/lib/crates/fabro-test/src/lib.rs
@@ -0,0 +1,290 @@
+use std::path::PathBuf;
+use std::process::Output;
+
+use assert_cmd::Command;
+use regex::Regex;
+
+/// Static filters applied to every snapshot.
+static INSTA_FILTERS: &[(&str, &str)] = &[
+    (r"fabro \d+\.\d+\.\d+", "fabro [VERSION]"),
+    (r"\b[0-9A-HJKMNP-TV-Z]{26}\b", "[ULID]"),
+    (r"in \d+(\.\d+)?(ms|s)", "in [TIME]"),
+    (r"\\([\w\d])", "/$1"),
+];
+
+/// An isolated test context for running fabro CLI commands.
+///
+/// Creates temporary directories for home, storage, and working directory,
+/// and provides methods to build commands with proper isolation env vars.
+pub struct TestContext {
+    pub temp_dir: PathBuf,
+    pub home_dir: PathBuf,
+    pub storage_dir: PathBuf,
+    fabro_bin: PathBuf,
+    filters: Vec<(String, String)>,
+    _root: tempfile::TempDir,
+}
+
+impl TestContext {
+    /// Create a new isolated test context.
+    ///
+    /// `fabro_bin` should be the path to the compiled `fabro` binary,
+    /// typically obtained via `env!("CARGO_BIN_EXE_fabro")`.
+    pub fn new(fabro_bin: PathBuf) -> Self {
+        let root = tempfile::tempdir().expect("failed to create temp dir");
+        let root_path = root.path().to_path_buf();
+
+        let temp_dir = root_path.join("temp");
+        let home_dir = root_path.join("home");
+        let storage_dir = root_path.join("storage");
+
+        std::fs::create_dir_all(&temp_dir).expect("failed to create temp_dir");
+        std::fs::create_dir_all(&home_dir).expect("failed to create home_dir");
+        std::fs::create_dir_all(&storage_dir).expect("failed to create storage_dir");
+
+        let filters = vec![
+            (
+                regex::escape(temp_dir.to_str().unwrap()),
+                "[TEMP_DIR]".to_string(),
+            ),
+            (
+                regex::escape(home_dir.to_str().unwrap()),
+                "[HOME_DIR]".to_string(),
+            ),
+            (
+                regex::escape(storage_dir.to_str().unwrap()),
+                "[STORAGE_DIR]".to_string(),
+            ),
+        ];
+
+        Self {
+            temp_dir,
+            home_dir,
+            storage_dir,
+            fabro_bin,
+            filters,
+            _root: root,
+        }
+    }
+
+    /// Returns the combined static + context-specific filters.
+    pub fn filters(&self) -> Vec<(String, String)> {
+        let mut filters: Vec<(String, String)> = INSTA_FILTERS
+            .iter()
+            .map(|(pat, rep)| ((*pat).to_string(), (*rep).to_string()))
+            .collect();
+        filters.extend(self.filters.clone());
+        filters
+    }
+
+    /// Build a base `Command` with all isolation env vars set.
+    pub fn command(&self) -> Command {
+        let mut cmd = Command::new(&self.fabro_bin);
+        cmd.env("NO_COLOR", "1");
+        cmd.env("HOME", &self.home_dir);
+        cmd.env("FABRO_NO_UPGRADE_CHECK", "true");
+        cmd.env("FABRO_STORAGE_DIR", &self.storage_dir);
+        cmd
+    }
+
+    /// Build a `validate` subcommand.
+    pub fn validate(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("validate");
+        cmd
+    }
+
+    /// Build a `run` subcommand.
+    pub fn run_cmd(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("run");
+        cmd
+    }
+
+    /// Build a `ps` subcommand.
+    pub fn ps(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("ps");
+        cmd
+    }
+
+    /// Build a `model` subcommand.
+    pub fn model(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("model");
+        cmd
+    }
+
+    /// Build a `secret` subcommand.
+    pub fn secret(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("secret");
+        cmd
+    }
+
+    /// Build a `doctor` subcommand.
+    pub fn doctor(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("doctor");
+        cmd
+    }
+
+    /// Build a `exec` subcommand.
+    pub fn exec_cmd(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("exec");
+        cmd
+    }
+
+    /// Build a `llm` subcommand.
+    pub fn llm(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("llm");
+        cmd
+    }
+
+    /// Build a `config` subcommand.
+    pub fn config(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("config");
+        cmd
+    }
+
+    /// Build a `cp` subcommand.
+    pub fn cp(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("cp");
+        cmd
+    }
+
+    /// Build an `init` subcommand.
+    pub fn init_cmd(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("init");
+        cmd
+    }
+
+    /// Build an `install` subcommand.
+    pub fn install(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("install");
+        cmd
+    }
+
+    /// Build a `pr` subcommand.
+    pub fn pr(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("pr");
+        cmd
+    }
+
+    /// Build a `preview` subcommand.
+    pub fn preview(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("preview");
+        cmd
+    }
+
+    /// Build a `repo` subcommand.
+    pub fn repo(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("repo");
+        cmd
+    }
+
+    /// Build an `ssh` subcommand.
+    pub fn ssh(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("ssh");
+        cmd
+    }
+
+    /// Build a `system` subcommand.
+    pub fn system(&self) -> Command {
+        let mut cmd = self.command();
+        cmd.arg("system");
+        cmd
+    }
+}
+
+/// Execute a command and format the output for snapshot testing.
+///
+/// Returns the formatted string and the raw `Output`.
+/// Prints unfiltered output to stderr for debugging failed tests.
+pub fn run_and_format(cmd: &mut Command, filters: &[(String, String)]) -> (String, Output) {
+    let output = cmd.output().expect("failed to execute command");
+
+    let stdout = String::from_utf8_lossy(&output.stdout).to_string();
+    let stderr = String::from_utf8_lossy(&output.stderr).to_string();
+
+    // Print unfiltered output for debugging
+    #[allow(clippy::print_stderr)]
+    {
+        eprint!("{stdout}");
+        eprint!("{stderr}");
+    }
+
+    let filtered_stdout = apply_filters(&stdout, filters);
+    let filtered_stderr = apply_filters(&stderr, filters);
+
+    let formatted = format!(
+        "success: {success}\nexit_code: {code}\n----- stdout -----\n{stdout}----- stderr -----\n{stderr}",
+        success = output.status.success(),
+        code = output.status.code().unwrap_or(-1),
+        stdout = filtered_stdout,
+        stderr = filtered_stderr,
+    );
+
+    (formatted, output)
+}
+
+/// Apply regex-based filters to a snapshot string.
+pub fn apply_filters(snapshot: &str, filters: &[(String, String)]) -> String {
+    let mut result = snapshot.to_string();
+    for (pattern, replacement) in filters {
+        if let Ok(re) = Regex::new(pattern) {
+            result = re.replace_all(&result, replacement.as_str()).to_string();
+        }
+    }
+    result
+}
+
+/// Create a `TestContext` using the `fabro` binary built by cargo.
+#[macro_export]
+macro_rules! test_context {
+    () => {
+        $crate::TestContext::new(std::path::PathBuf::from(env!("CARGO_BIN_EXE_fabro")))
+    };
+}
+
+/// Snapshot test macro that runs a command and compares output using insta.
+///
+/// Usage:
+/// ```ignore
+/// fabro_snapshot!(context.filters(), context.validate().arg("--help"), @"...");
+/// ```
+#[macro_export]
+macro_rules! fabro_snapshot {
+    ($spawnable:expr, @$snapshot:literal) => {{
+        let filters: Vec<(String, String)> = $crate::TestContext::default_filters();
+        let mut cmd = $spawnable;
+        let (snapshot, _output) = $crate::run_and_format(&mut cmd, &filters);
+        insta::assert_snapshot!(snapshot, @$snapshot);
+    }};
+    ($filters:expr, $spawnable:expr, @$snapshot:literal) => {{
+        let filters: Vec<(String, String)> = $filters;
+        let mut cmd = $spawnable;
+        let (snapshot, _output) = $crate::run_and_format(&mut cmd, &filters);
+        insta::assert_snapshot!(snapshot, @$snapshot);
+    }};
+}
+
+impl TestContext {
+    /// Returns just the static default filters (no context-specific paths).
+    pub fn default_filters() -> Vec<(String, String)> {
+        INSTA_FILTERS
+            .iter()
+            .map(|(pat, rep)| ((*pat).to_string(), (*rep).to_string()))
+            .collect()
+    }
+}