mirror of
https://github.com/abhigyanpatwari/GitNexus.git
synced 2026-09-17 23:52:36 +00:00
b1d87c1f33
* fix(eval): cut skill-evolution wall clock without shrinking the gate
Reuse matching incumbent/CE cells, sanitize each SHA once, and default
dispatch workers to 3 so weekly review generations finish inside the
EventBridge window. Cap the sweep from leftover instance uptime so a
Friday dispatch still uploads evidence.
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* perf(eval): pipeline graph setup and correct the wall-clock cost model
The evolution sweep paid `sanitize` + `analyze --pdg --index-only` for every
unique task SHA on the critical path, one at a time, with nothing overlapping.
`_run_sweep` now starts the next unpaid SHA's clone template and graph snapshot
on a prefetch thread as soon as the current task's cells are dispatched, so
every SHA but the first hides behind a paid session wave. The thread is joined
before that SHA is used and before the trees tempdir is torn down, and a
prefetch failure is recorded against the SHA exactly as an inline failure is.
Tasks whose cells are all reusable comparator rows are not prefetched: they
never build a graph, so priming one would be pure cost.
Adds `measure_evolution_cost.py`, the cost model behind these numbers. It reads
the review corpus, the evolve defaults, and the workflow's workers default —
it does not start a session. Its first version charged `copy_isolated_tree`
once per paid cell, serially. `run_cell` clones inside its own pool worker, so
the clones in a wave overlap and only one is on the critical path per wave;
the model now charges `ceil(cells / workers)` waves.
Estimated review generation at workers=3: cold 21570s, weekly 7710s.
Wall clock is quantised by `ceil(cells_per_task / workers)`. A cold review task
is 9 cells, so workers=4 buys the wall clock of workers=3 and pays host
contention for it. Documented in the workflow's rollout checklist.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* perf(eval): price the benchmark against measured cell durations
The cost model assumed every cell runs the 1140s mean. Cells are not uniform:
the 41 rows in Actions run 33912693948's artifact are 826s at the median,
1262s at the mean, 2976s at p90, with two pinned at the 5400s session ceiling.
A wave waits for its slowest cell, so a mean understates every concurrent
schedule — the previous model called workers=3 cold 5.99h when the same
schedule against real durations is 10.33h.
session_durations.json carries the sample in submission order with its
provenance and its caveat: every cell in that run returned unusable evidence,
so the durations are real but a clean run may sit lower. It is the only live
artifact; the 2026-07-22 green run's has expired.
The model now simulates the schedule cell by cell rather than multiplying a
mean by a wave count, averaged over all 41 rotations of the sample so no
single alignment between sample order and cell index decides the answer. It
prices today's barrier (wave_makespan) against a continuously fed pool
(fed_makespan) and reports both, and it charges the proposer session — one
per generation, measured at 344.7s — which it had been omitting entirely.
Measurement only; no runtime behaviour changes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* perf(eval): price arms separately and stop inventing setup constants
Two errors in the model, both found by auditing it against the artifact it
claims to describe.
The arms are not interchangeable. `candidate_review` runs 1416s at the mean
against `review`'s 1204s and `ce_review`'s 1176s, and the weekly lane pays the
candidate arm and nothing else — reuse skips both incumbents. Pricing weekly
from a pooled sample charged it for arms it never runs: weekly is 4.59h, not
the 3.65h a pooled sample reported. Cells are also submitted run-major and
arm-minor, so at workers=3 every wave holds one cell of each arm and the
slowest arm sets the wave; the model now builds cells in that order.
The setup constants were invented. GRAPH_ANALYZE_SECONDS=600 and
TEMPLATE_SANITIZE_SECONDS=180 charged 3900s of per-SHA setup for a cold run —
more than the entire non-session time of the source run, which was 2541s for
41 cells and 5 SHAs. `duration_s` is the sum of a cell's Claude sessions
(runner_sessions.py), so that 2541s residual is every clone, graph build,
sandbox and teardown the sweep paid. The model now charges the measured
residual per cell, 62.0s, and no longer credits clone templates or graph
prefetch: both landed after that run and there is no measurement of them yet.
The residual bounds what they can be worth.
Cold 37452s (10.40h), weekly 16541s (4.59h), against a fed pool at 31683s and
16541s. Measurement only; no runtime behaviour changes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* perf(eval): charge sweep overhead where more workers cannot dissolve it
Three defects, found by auditing the model against the artifact again.
The overhead was charged inside the schedule. session_durations.json claimed
the residual was charged "per cell and serially - the pessimistic reading",
but task_cells folded it into each cell's duration, where the pool then
divided it by the worker count. The residual mixes per-cell work the pool
really does divide with per-SHA graph setup it cannot, and the artifact cannot
separate them, so it now sits outside the schedule: cold 11.09h, not 10.40h.
Alignment averaging weighted the shortest sample twice. The arm samples are 13,
14 and 14 long and the average ran over max()=14 offsets, so candidate_review's
first cell was counted twice and its last never. Averaging over lcm()=182
offsets weights every arm's sample evenly.
The wall assumed all 54 cells run. Replaying the sample's own error_kind
sequence through today's systemic_outage_streak trips the outage breaker at
cell 5 of 41. The source run executed all 41, so its runner did not break on
that sequence, but the current one would: these numbers price a HEALTHY sweep,
and a sweep with the sample's failure profile never reaches them. Stated on
generation_seconds and recorded next to the sample it qualifies.
Measurement only; no runtime behaviour changes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(eval): give the review agent somewhere it can actually write
Every review cell in the last recorded generation returned unusable evidence.
Not some — all 41, across all three arms and all six tasks, at $3653 for the
run. The transcripts say why, 127 times across 35 of 35 sessions:
EROFS: read-only file system,
open '/workspace/review-output.json.tmp.2.90a76e583b0c'
The review arm mounted the artifact as a writable FILE at
/workspace/review-output.json while binding /workspace read-only. The Write
tool writes atomically: it creates `<target>.tmp.<n>.<hex>` beside the target
and renames it. The parent was read-only, so the temp create failed and the
artifact was never written. A writable file inside a read-only directory is
not writable to anything that writes atomically. Agents tried
/proc/self/root/workspace/... and /proc/1/root/workspace/... to get around it;
all 41 artifacts came back 0 bytes.
The artifact now lives in its own writable directory bound at /review-output,
outside the workspace. That is what a rename needs, and it lets the workspace
get stricter rather than looser: the review phase may now change nothing there
at all (enforce_phase_workspace gained allowed_artifact=None), where before it
was entitled to one path inside it. The file is no longer pre-created — the
agent writes it, and absence is now meaningful evidence.
parse_review_output reported every one of these as "review output is not valid
UTF-8 JSON". The file was empty, and its except folded OSError, UnicodeError
and JSONDecodeError into that one string, so a sandbox that made writing
impossible was indistinguishable from an encoding fault. That is why this read
as an agent-quality problem for fifteen consecutive non-green runs. Each cause
now names itself: never written, empty, not valid UTF-8, not valid JSON with
the decoder's position. run_arm also keeps the FIRST error_detail, as it
already did for error_kind, so a phase-boundary violation is no longer buried
under the parse failure it causes.
The test double conflated sandbox.private_root with the clone, which put the
artifact directory inside the workspace and would have hidden the stricter
check. Regression tests pin the mount shape in the generated bwrap argv, the
contract path in the prompt, the four parse diagnostics, and the
untouched-workspace contract.
Verified by unit tests only: this container has unprivileged user namespaces
disabled, so bwrap cannot run here and the mount was not exercised end to end.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(review): close the artifact path in every layer that gates it
Code review of this branch found the relocated review artifact was fixed in the
bwrap mount and nowhere else. Four independent layers decide whether the agent
can write it, and three still named the old location.
Claude Code applies its own filesystem policy to its own tools, and
build_claude_settings listed only /workspace, /tmp and /home/agent under
allowWrite with denyRead ["/"]. The artifact used to live under /workspace, so
this list was correct until it moved. SANDBOX_REVIEW_OUTPUT is now in allowWrite
and allowRead; without it the bwrap bind grants a write the CLI then refuses.
The task corpus still ran `test -s review-output.json` from the workspace, in a
separate sandbox invocation that never sees the artifact mount. Every review
cell would have been stamped verify-failed with resolved=False no matter how
good the review was, which also made those rows permanently unreusable and so
silently disabled this branch's own comparator reuse for review arms. The verify
and hidden-oracle commands now read the location from
GITNEXUS_BENCH_REVIEW_OUTPUT and get the directory bound read-only, mirroring
the mount-plus-env-var shape _run_hidden_oracle already used.
host_text and host_path did not translate the new path, so the host-unsafe
backend told the agent to write somewhere that exists on neither backend.
Adding the mapping exposed a second defect: host_text substituted every
occurrence of a target, and "/review-output" appears twice in
"/review-output/review-output.json" - once as the directory and once inside the
filename. Matching is now anchored to a path boundary.
Comparator reuse had three ways to accept evidence it should have rejected. A
row with no runtime_digest passed the drift lock because the guard only compared
when both sides were bound, and the branch's own test asserted that as correct;
absence is now a mismatch and the test states the rule. materialize_reused_row
overwrote recorded_at with the copy time while the age check read that field, so
a row copied forward each generation refreshed its own clock and never aged out;
the first measurement time is now preserved and aged against. A future-dated
stamp passed a one-sided bound and is now rejected as corrupt.
RUNTIME_DIGEST never reached the runner at all: runner_environment builds a
fixed dict and process_control replaces the child environment wholesale, so the
digest the workflow exports was dropped and the lock it feeds was inert. The
instance-window deadline was also checked only after run_proposer returned,
buying a proposal the generation had no room to benchmark.
The graph prefetch thread was started without copy_context, so it never saw the
cancellation ContextVar the rest of the sweep shares, and the outage breaker
returned without setting cancel_event - together, a tripped breaker would block
on joining a prefetch that was never told to stop. Both fixed, with outage
checked before cancellation at the two exits so an outage keeps exit 1 instead
of becoming a Ctrl-C's 130.
Both bwrap canaries that actually execute a write still bound the pre-fix shape
against a file this branch no longer creates, so they would have errored rather
than caught anything. They now bind the directory and write atomically - temp
file beside the target, then rename - which is the exact operation that failed
with EROFS. A source-text assertion over inspect.getsource(run_arm) was replaced
with one that inspects the real mount, and a wall-clock assertion was pinned to
a fixed monotonic clock.
Not applied, and why: binding task-asset and dependency digests into comparator
reuse needs asset snapshots prepared before the reuse decision rather than
inside the per-task loop, and shipping the comparison without that would add a
guard that silently never fires. Forcing a paid canary cell per incumbent arm
and folding reused rows into the outage streak are behaviour decisions, not
fixes. Clone-template reuse still has no test. The cost model's per-cell
residual still shrinks with arm count, overstating weekly savings by at most the
2541s residual; the docstring now says so rather than inventing a split.
585 eval tests pass, ruff clean, 29 workflow contract tests pass. The two
test_model_gateway.py failures are pre-existing and fail on main.
* fix(review): bind reuse to its environment and keep the health canary real
Applies the five findings the previous review round left open.
Comparator reuse ignored the environment a row was measured in. TaskReuseBinding
carried the task and oracle identity but not the task-asset or sandbox-dependency
digests, and this branch itself changes sandbox_dependencies in the review
corpus - so a reused comparator could be measured against one dependency set and
compared against a candidate built on another, handing the gate a false
comparison. Closing it needed the digests to exist before the reuse decision, so
asset snapshots are now prepared for every task up front instead of lazily
inside the per-task loop. That also removes the concurrent TaskAssetCache.prepare
the prefetch thread could otherwise race, which the file's own "plain dict,
read-then-write race" comment warned about. Both digests fail closed on either
side, matching the runtime digest.
The broken-incumbent canary could not fire when reuse was working. It read
`resolved`, which counts reused rows, so an arm whose cells were all reused
always looked healthy - in precisely the run where a broken environment would go
unnoticed. aggregate now also reports `resolved_fresh` and the canary reads it.
That count would be vacuous if an arm were reused end to end, so the sweep keeps
one paid cell per incumbent arm and says which one it kept.
Reused rows did not participate in the outage streak, so a run of failures could
carry across them and trip on stale history. A reused success now resets the
streak the way a paid success does.
The cost model charged sweep overhead per cell, which credited a weekly
generation for shrinking work it still performs: it pays one arm instead of
three but builds exactly the same graphs. Overhead is charged per SHA now.
Weekly is 5.20h rather than the 4.80h the per-cell rate reported; cold is
10.86h. The residual still cannot be split between per-SHA and per-cell work
from one artifact, so session_durations.json records that assumption and the
direction it errs in, rather than leaving a number nobody can trace.
Clone-template reuse - the branch's core speedup, taken on essentially every
multi-cell sweep - now has a test that builds a real sanitized template, asserts
the cell runs against the copy with the template's HEAD, and fails if run_cell
re-clones. A second test asserting only on a namespace built inside the test was
written and deleted: it exercised nothing, which is the failure this review
round penalised elsewhere.
589 eval tests pass, ruff clean, 29 workflow contract tests pass. The two
test_model_gateway.py failures are pre-existing and fail on main.
* refactor(eval): consolidate duplicated harness logic after the review round
Simplification pass over the branch. Behavior-preserving throughout; three
reviewers, nine findings applied, two skipped.
The review-artifact block in _run_hidden_oracle was unreachable. That function
runs only in run_arm's non-review branch, while the directory it probes for is
created only in the review branch, and each sandbox serves exactly one arm - so
`review_artifact.parent.is_dir()` could never be true. It was added an hour
earlier to make the hidden oracle resolve the moved artifact; the oracle never
runs for review tasks, so the guard was dead on arrival. Deleting it also
removes the duplication it had with the verify-command wiring.
EXCLUDED_ERROR_KINDS is now one definition. runner.py and comparator_reuse.py
each carried the same six-member frozenset, kept in sync by a comment. Only one
direction is possible: runner already imports from comparator_reuse, so the
reverse import fails at module-init with a circular-import error. That is now
stated where the alias lives, so nobody tries it the other way.
ensure_task_graph and prefetch_next_graph shared ten keyword parameters, passed
through two call sites and forwarded whole between them. They now take a
GraphBuildEnv, mirroring TaskCellContext, which already bundles per-cell state
in this file. Its ready_keys() replaces an inline four-set union at the call
site.
Smaller consolidations: _sha256_file's hand-rolled chunk loop becomes
hashlib.file_digest (3.11+, already used in runner_artifacts); _copy_owner_only
reuses task_assets._write_all and COPY_CHUNK_BYTES instead of repeating the
short-write retry; its stat-then-open existence check becomes the O_EXCL failure
it was already relying on, which is atomic rather than merely narrow; and
runner_environment reads the digest through comparator_reuse.current_runtime_digest
instead of re-parsing the environment variable.
Three test docstrings summarised the branch's own history ("the branch's core
speedup", "the regression that produced fifteen runs") rather than the invariant
under test. Rewritten to state the constraint, which is what survives the merge.
Repaired the indentation left behind by the outage-streak edit and flattened the
prefetch dispatch from three nested conditionals to one.
Skipped: consolidating comparator_reuse._real_directory onto proposer_sandbox's
same-named helper - they differ, the sandbox one rejects any symlink in the
resolved path while this one checks only the leaf, so sharing it would tighten
behavior rather than preserve it. That needs a decision about which policy the
reuse path wants, not a simplification.
589 eval tests pass, ruff clean, 29 workflow contract tests pass. Unrelated and
pre-existing: two test_model_gateway.py failures, and
test_process_control.py::test_timeout_kills_term_ignoring_descendants_before_they_write,
which is a TERM-to-KILL timing flake (passes 2 of 3 in isolation) in a file this
branch does not touch.
* refactor(eval): name the reuse directory check for the promise it makes
The simplification pass left one finding open: comparator_reuse and
proposer_sandbox both defined `_real_directory`, same name and same shape, with
different guarantees. The sandbox one rejects every symlink hop in the path; the
reuse one checks only the leaf and resolves through parents. Sharing the name
invites a consolidation that would silently tighten one of them.
They should not be merged, so the name stops claiming they could be.
proposer_sandbox guards a mount root, where a symlink hop changes what an
untrusted session is handed. comparator_reuse guards a data directory whose
contents are already validated one file at a time - reads go through
_regular_file, which lstats and rejects symlinks, and writes through O_NOFOLLOW.
A symlinked parent therefore grants nothing those guards do not already cover,
while refusing one would reject a symlinked artifacts directory or macOS's /var
for no gain.
Renamed to _resolved_directory, with the reasoning recorded at the definition,
and a test that pins both halves: a symlinked parent is accepted and resolved, a
symlinked leaf is still refused. Behavior is unchanged.
591 eval tests pass, ruff clean. The two test_model_gateway.py failures are
pre-existing and fail on main.
* test(eval): measure the sweep scheduler instead of modelling it
measure_evolution_cost predicts wall clock from a model of what
sweep_task_cells does. This runs the real thing - real threads, the real wave
barrier, the real outage breaker - with only the paid agent session replaced by
a sleep, and times it.
Durations are the measured per-arm samples divided by 5000, so a 1416s cell
takes ~0.28s. The shape is kept on purpose: the median cell is 826s against a
5400s ceiling, and that spread is the entire reason a barrier costs anything.
Uniform random sleeps would erase the effect under test. All schedulers consume
one identical seeded plan, so a comparison cannot be an artifact of one of them
drawing luckier cells.
The model survives contact: it tracks real execution within about 10%, and
workers=1 - which runs without a pool at all - sits at 0.95, so the residual
above 1.0 at higher worker counts is per-wave thread overhead rather than a
modelling error. Two structural claims that were arithmetic are now observed.
Weekly is flat from workers=3: 3.59, 3.59, 3.59, 3.60, 3.59, 3.59 across w=3..8.
workers=4 buys nothing over workers=3 on cold, 7.68 against 7.78.
Two prototype schedulers are measured beside it, deliberately before any
production code exists. A continuously fed pool per task is worth more than the
model claimed on cold, -27.3% against a predicted -17.9%, and exactly nothing on
weekly, +0.0%, because a weekly task is one wave with nothing to feed. One pool
across all tasks beats both: -40.7% weekly and -42.9% cold at workers=3, rising
to -65.7% and -63.9% at workers=8. It also subsumes the fed pool, since packing
across tasks is a fed pool.
That reorders the backlog. Cross-task packing moves from second to first: it
dominates on both profiles, and it is the only thing that moves weekly at all.
Raising the worker count is worth nothing until it lands - under the barrier
weekly does not improve from w=3 to w=8, and speedup against serial is 1.58x for
three workers and only 2.40x for eight.
The bound on all of it: sleeping threads do not contend. Real sandboxed sessions
compete for CPU, page cache and disk, and the duration sample was itself
measured at workers=1, so it carries no contention either. These speedups are
upper bounds. The ordering is trustworthy because the schedulers were compared
under identical conditions; the magnitudes are not. The packed prototype is also
a bare ThreadPoolExecutor with no breaker folding, no per-task graph lifecycle
and no reuse binding - which is the actual cost of building it, and is not
measured here.
* test(eval): carry the sweep invariants into the packed prototype
The first packed prototype was a bare ThreadPoolExecutor. It reported -43% and
none of the invariants the shipped scheduler holds, so it priced an idea nobody
could ship. This one carries them: a global submission order continued across
task boundaries, in-order folding, the real outage breaker, and per-task graph
readiness gating behind a serial builder.
The fidelity check first reported the two schedulers tripping on different
cells, 17 against 16. That was my instrumentation, not a divergence -
sweep_task_cells folds an entire wave before it evaluates the breaker, so the
last cell folded is not the cell that tripped. With the harness mirroring the
breaker's own evaluation the two agree exactly, across failures starting at
cell 0, 4 and 12, with overrun inside the workers-1 bound the wave docstring
promises.
Two results worth the exercise.
Head-of-line blocking, not the barrier, is what a naive in-order design pays.
Holding submission to `workers` cells beyond the fold pointer leaves the
faithful scheduler at -8.1% cold and -2.7% weekly: one slow cell stalls the
pointer, the window cannot slide, and it reproduces the wave almost exactly.
That is the number to quote if anyone proposes the obvious implementation.
But the overrun bound turns out to be set by the worker count, not the window.
Only `workers` cells can be running when the breaker trips; everything queued
behind them short-circuits on the halt flag. Overrun is 3 at an unbounded
window exactly as at 6, and the trip cell never moves off 16. So H2 does not
have to trade breaker fidelity for speed - a wide window takes -42% with the
semantics intact. The tension I assumed was there is not, and window=12 already
captures 97% of it.
Still an upper bound: sleeping threads do not contend, and the sample was
measured at workers=1. What this establishes is that the invariants are
affordable, which was the thing blocking H2. Not built here: the trees tempdir
lifecycle, reuse-row binding, and the cancel_event path.
591 eval tests pass, ruff clean.
* test(eval): put the scheduler comparison under real CPU contention
Every Phase 2 number so far came from sleeping threads, which contend for
nothing, against a duration sample measured at workers=1, which contains no
contention either. That was the standing caveat on the whole result, so this
measures it.
A cell now waits for its API share and then burns a fixed number of sha256
rounds in a subprocess. Work-bounded rather than wall-clock bounded, so it takes
longer when cores are busy - that is the effect under test. A subprocess because
Python threads burning Python would measure the GIL rather than the machine.
Calibrated at 519k rounds/s, stable within 2% across three probes.
The first run of this was worthless and is recorded as such: on a 24-core host
with 3 to 6 workers nothing ever contends, since cpu_fraction 0.5 at 6 workers
is about 3 cores of demand out of 24. It measured an absence. Re-run pinned with
taskset to 4 and 2 cores.
The packing advantage survives. It holds between -40% and -47% across every host
size and CPU fraction tested, including a genuinely oversubscribed 2-core box at
cpu_fraction 0.5 with 6 workers.
But contention erodes packing more than it erodes waves, for a structural
reason: packing is what creates the concurrency. Moving from 24 cores to 2 at
cpu 0.5 and 6 workers, the faithful scheduler slows 13% while the wave slows
3.7%, and the gain narrows from 45.0% to 39.8%. Packing and a higher worker
count are therefore not independent wins - packing spends the contention
headroom first, so raising workers has to be re-argued after it lands rather
than added to it.
Three things this still does not measure, and they bound the result. The real
CPU fraction of a benchmark cell is a guess informed by roughly 180 tool calls
per session; nobody has profiled one. The evolution runner's core count decides
which column applies and is unknown here. And the burn is sha256, pure CPU,
while real cells run vitest and analyze, which are memory and IO heavy - so this
is a floor on contention, not a ceiling.
591 eval tests pass, ruff clean.
* perf(eval): add a packed sweep scheduler, and correct the bound I claimed for it
sweep_task_cells finishes one task before starting the next and drains a wave
before refilling it, so a task with fewer cells than workers leaves workers
idle and one slow cell stalls its whole wave. sweep_packed_cells feeds every
task's cells through a single pool instead. Measured against the review corpus
it is worth about 40% of a cold sweep, and it is the only change that moves a
seeded weekly run at all - there a task is three cells and a wave is never full.
The breaker keeps its exact meaning. Cells carry a total submission order
continued across task boundaries, a folder walks results in that order, and
consecutive systemic failures are counted there, so a doomed run aborts on the
same cell it would have under waves. Verified at three failure positions.
This commit also corrects a finding from the Phase 2 prototype. I claimed the
overrun bound was set by the worker count rather than the submission window,
and that packing therefore cost nothing in breaker fidelity. That was derived
from a window sweep that only ever injected failures at one position. Driving
the real function at other positions shows the halt flag does not bound overrun
at all: the folder walks in order, so a slow early cell lets workers race ahead
and the trip is detected after those cells have already paid. An unbounded
queue overran by 11 cells where waves overrun by 2.
So the window is load-bearing and the trade is real, measured at workers=3 with
failures injected at four positions:
window 3 -> -8% wall, overrun 2 (the wave scheduler's own bound)
window 6 -> -27% wall, overrun 4
window 12 -> -42% wall, overrun 9
window 54 -> -44% wall, overrun 11
Overrun is wasted paid sessions at roughly $70 each. The default multiplier is
2, keeping the worst case within twice the wave bound while taking most of the
gain; the curve is in the constant's comment so raising it is an informed
decision rather than a guess.
Not wired in yet: _run_sweep still calls sweep_task_cells per task. Moving the
per-task graph, trees tempdir and reuse binding out of that loop behind
await_ready is the larger and riskier half, and it belongs in its own change.
595 eval tests pass, ruff clean.
* fix(eval): judge harness health on execution, not on how many tasks resolved
broken_incumbent_arms infers "the environment is broken" from an arm resolving
zero tasks. That inference does not hold: a reviewer can be wrong about every
task in a hard corpus while every process, mount and capture worked perfectly.
Actions run 33962002890 is exactly that shape - 51 cells, all resolved=False
with error_kind=oracle-failed, median score 0.212, and a healthy harness.
Someone already knew this, and patched it by excluding review arms at the call
site. That leaves the unsound inference in place for workflow and
workflow_direct, and leaves review arms with no health check at all - so the
run that genuinely was broken, 33912693948, where the mount made an atomic
write impossible and all 41 artifacts came back empty, could not have been
caught here either.
So this replaces the inference rather than adding another exemption. aggregate
now classifies fresh rows into execution failures (the process or its tooling
did not complete), evidence failures (it completed but produced nothing
trustworthy or scoreable), and admissible measurements. An arm is unhealthy
only when it has fresh attempts, zero admissible measurements, and at least one
execution or evidence failure. Resolution count is no longer consulted. Arms
with only reused rows report current health as UNKNOWN rather than good.
With the inference corrected, review arms are checked again, which is what lets
the empty-artifact case be caught at all.
Deliberately unchanged: comparator reuse eligibility, quality denominators,
promotion thresholds, model settings, skill prompts and scheduler behaviour.
Failures that stop being called infrastructure failures still surface in the
counts and reasons - an agent-originated failure must not vanish from reporting
because it was reclassified. broken_incumbent_arms and its tests are left in
place; deleting behaviour belongs in its own change.
Seven regression tests, built from both runs' shapes and labelled as
reconstructed from logged observations, since 33962002890's results.jsonl did
not survive the instance shutdown. They pin: a badly-scoring reviewer is
healthy; an all-zero score is still a valid negative; empty artifacts are
unhealthy; one admissible cell keeps an arm healthy while its failures stay
visible; reused rows alone leave health unknown; reused successes do not mask
fresh failures; and a parseable artifact does not excuse a failed session.
602 eval tests pass, ruff clean.
* fix(eval): pin the health guard below the breaker, and stop calling mixed runs healthy
Two corrections to the health-classification patch.
The regression I wrote could not have proved what it claimed. A fixture of 41
empty artifacts aborts through the outage breaker long before finalization:
review-evidence-invalid is in SYSTEMIC_ERROR_KINDS and the limit is 5, so it
trips at cell 5 through the pre-existing path. It demonstrated failure
detection, not the new guard. The decisive test now uses ONE fresh unusable
cell, asserts the streak stays under the breaker threshold, and only then
requires finalization to abort - leaving the new check as the only thing that
can catch it. Removing the call makes that test fail; restoring it passes.
The accurate defect statement is narrower than the last message claimed. Review
arms were excluded from the final incumbent-health check while the consecutive-
failure breaker gave them separate, partial coverage. They were not unguarded.
Second: "one admissible cell plus two execution failures" was asserted as
healthy. That converts "not wholly unusable" into "ran reliably", which is how
a partly-broken sweep passes review. Arms now report UNKNOWN, OBSERVED_OK,
DEGRADED or UNUSABLE. Only UNUSABLE is fatal, so eligibility and promotion are
untouched - this changes what is reported, not what is allowed.
The guard is extracted as enforce_measurement_health so it can be driven
directly, and it now reports a status line per arm. It names no cause: an empty
artifact establishes that evidence is unusable, not that a mount rejected the
write, so it prints cause=undetermined rather than guessing EROFS. It still
runs after report.md and promotion.json are written, so a failing sweep leaves
its evidence behind.
ce_review is named explicitly at the call site. It is a comparator rather than
a candidate, so it is absent from CANDIDATE_ARMS.values(), and dropping the
review exclusion alone would have left it unclassified.
The wiring test reads _run_sweep's compiled code object for the referenced
global rather than matching source text. It is honest about its limit: it
proves the call exists and would catch its removal, but no test here drives
_run_sweep end to end, which needs bwrap and a sandbox.
broken_incumbent_arms is marked LEGACY and NON-AUTHORITATIVE with removal
tracked. It has no production caller.
608 eval tests pass, ruff clean. The two test_model_gateway.py failures are
test_locked_litellm_translates_messages_to_offline_responses and
test_openai_gateway_never_leaves_proxy_output_on_an_undrained_pipe; both fail
identically on origin/main in this environment, checked directly rather than
carried forward as an inherited label.
* fix(eval): review artifact path, evidence classification, comparator reuse
Extracted from the combined skill-evolution branch. This is the runtime change
set: everything that alters how a sweep executes and what it records. The
packed scheduler and its measurement harness were separated onto
perf/skill-evolution-packed-scheduler, which is purely additive.
Correctness. The review artifact was mounted as a writable FILE inside a
read-only workspace while the agent's Write tool writes atomically - temp file
beside the target, then rename - so the temp create failed EROFS and the
artifact was never written. Four layers gate that path and three named the old
location: the CLI's own allowWrite/allowRead policy, the task corpus verify
command run in its own sandbox invocation, and host_text/host_path for the
host-unsafe backend. Fixing the translator exposed a second defect, since
"/review-output" appears twice in "/review-output/review-output.json"; matching
is now anchored to a path boundary. parse_review_output folded OSError,
UnicodeError and JSONDecodeError into one message, so an artifact that was
never written looked like an encoding fault; each cause now names itself.
Health classification. broken_incumbent_arms inferred a broken environment from
an arm resolving zero tasks, which a reviewer facing a hard corpus falsifies -
Actions run 33962002890 is exactly that shape. Arms are now classified from
fresh execution and evidence outcomes as UNKNOWN, OBSERVED_OK, DEGRADED or
UNUSABLE, and only UNUSABLE aborts. Resolution count is not consulted. The
guard names no cause: an empty artifact establishes unusable evidence, not that
a mount rejected the write.
Comparator reuse. Reuse accepted evidence it should have rejected: a row
without a runtime_digest passed the drift lock, recorded_at was overwritten with
the copy time so a row could outlive its own max_age, and the binding ignored
task-asset and dependency digests although this change alters
sandbox_dependencies in the review corpus. Closing the last one required
preparing asset snapshots before the reuse decision, which also removes the
concurrent TaskAssetCache.prepare the prefetch thread could race.
These three concerns share aggregate() and _run_sweep, which is why they ship
together: separating them further would mean hunk-level surgery on a function
all three modify, and the risk of a silent omission outweighs the reviewability
gain.
592 eval tests pass at this base. The two test_model_gateway.py failures,
test_locked_litellm_translates_messages_to_offline_responses and
test_openai_gateway_never_leaves_proxy_output_on_an_undrained_pipe, fail
identically on origin/main in this environment.
Known gap, and the reason this is not ready to merge: no test drives _run_sweep
end to end. enforce_measurement_health is unit-tested including the
below-breaker unusable case, and the caller wiring is pinned structurally by
reading _run_sweep's compiled code object, but interruption semantics, exit
precedence and persisted artifacts are not exercised through the real path.
* fix(eval): address PR review feedback (#3207)
- aggregate: count admissible rows directly instead of subtracting the
execution and evidence counters, which double-charged a row that is both
a session error and invalid review evidence and could report UNUSABLE for
an arm holding real measurements.
- run_proposer: bound the session timeout by what is left of
--max-runtime-seconds, so clearing the sweep minimum cannot start a
full-length session past the instance window.
- comparator reuse: hold one O_NOFOLLOW descriptor for the size check,
digest and copy, and prove it is the inode that was checked, closing the
swap window a concurrent writer of the reuse directory had.
- Drive the review-artifact mount assertion through run_arm and the
clone-template assertion through run_cell, instead of rebuilding the
expected values in the tests (also removes the CodeQL unnecessary lambda).
- Assert the workflow invokes run-evolution.sh rather than that its YAML
mentions --max-runtime-seconds, which only appears in a comment.
- Correct the parse_review_output failure-mode claim: the fold was empty
artifacts reported as "not valid UTF-8 JSON"; a never-created file raised
FileNotFoundError.
- prettier: wrap the over-long readFileSync call flagged by PR autofix.
Note: pre-existing failure in tests/test_model_gateway.py::test_locked_litellm_translates_messages_to_offline_responses (local LiteLLM proxy never becomes ready in this environment) not addressed by this PR.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(eval): address the second round of PR review feedback (#3207)
- Refuse a symlinked `transcripts` component on both sides of comparator
reuse. O_NOFOLLOW guards the leaf only, so a link there redirected the
read or the copy out of the results directory; checked per component as
evolution._require_directory_chain does.
- Base the paid incumbent canary on the cells this sweep PLANS. Reuse
selection accepts any prior run index, so a results directory produced
with more runs left extra keys, the equality never held, and the canary
stopped firing. Extracted as drop_canary_reuse_key and unit-tested.
- Start the runtime clock in main(). --max-runtime-seconds is measured from
/proc/uptime before exec, so parsing, task I/O, preflight and gateway
setup were being handed back to the sweep out of the upload reserve.
- Do not fall back to shutil.copytree when the managed clone copy was
cancelled or timed out; that fallback is for a filesystem that cannot
reflink, and copytree cannot be cancelled.
- Assert the review session's writable mount, not only the verifier's
read-only one: the EROFS bug is about the agent's write.
- Exercise ref isolation in the copy_isolated_tree test rather than
comparing an initial HEAD a shared namespace would also match.
- Point the stale-symlink fixture at the sentinel via os.path.relpath, and
skip the reuse symlink tests where symlink creation needs privilege.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(eval): close the runtime-cap gap and pin the reuse directory
Both were left open on #3207 as approach decisions rather than nits.
Runtime cap: run-evolution.sh computed the budget in its own
`uv run python -c` and passed a number, so the script's remaining
provenance work and the CLI's own startup were spent by nobody and charged
to the sweep — out of the upload reserve the cap exists to protect. The
script now passes --max-runtime-from-instance-window and evolve reads
/proc/uptime itself, on the line after it starts the clock the budget is
measured against, so no interval exists to lose. Also removes an
interpreter start from the script and lets --dry-run print the real argv.
Reuse directory: _real_child_directory lstat-checked `transcripts` and
returned its pathname, so a concurrent writer could rename the directory
and leave a symlink before the name was used again — O_NOFOLLOW guards
only the leaf. Every artifact is now resolved against a held descriptor:
_open_real_directory opens with O_DIRECTORY|O_NOFOLLOW (check and open in
one syscall), and _open_regular / _copy_owner_only take dir_fd. The reuse
path is therefore POSIX-only; _require_openat says so and fails closed,
which the runner already treats as "run a paid cell". _resolved_directory
still tolerates a symlinked reuse root, unchanged and still tested.
evolution._require_directory_chain is still lstat-per-component. It guards
a different surface (candidate overlay reads) that neither review raised,
so it is left alone rather than widened into here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(eval): sample the proposer budget where it is spent, digest what is copied
Four findings against
|
||
|---|---|---|
| .. | ||
| agents | ||
| analysis | ||
| bridge | ||
| configs | ||
| environments | ||
| prompts | ||
| tests | ||
| utils | ||
| workflow_bench | ||
| .env.example | ||
| .gitignore | ||
| __init__.py | ||
| constants.py | ||
| pyproject.toml | ||
| README.md | ||
| run_eval.py | ||
| tool_registry.py | ||
| uv.lock | ||
GitNexus SWE-bench Evaluation Harness
Evaluate whether GitNexus code intelligence improves AI agent performance on real software engineering tasks. Runs SWE-bench instances across multiple models and compares baseline (no graph) vs GitNexus-enhanced configurations.
What This Tests
Hypothesis: Giving AI agents structural code intelligence (call graphs, execution flows, blast radius analysis) improves their ability to resolve real GitHub issues — measured by resolve rate, cost, and efficiency.
Evaluation modes:
| Mode | What the agent gets |
|---|---|
baseline |
Standard bash tools (grep, find, cat, sed) — control group |
native |
Baseline + explicit GitNexus tools via eval-server (~100ms) |
native_augment |
Native tools + grep results automatically enriched with graph context (recommended) |
Recommended: Use
native_augmentmode. It mirrors the Claude Code model — the agent gets both explicit GitNexus tools (fast bash commands) AND automatic enrichment of grep results with callers, callees, and execution flows. The agent decides when to use explicit tools vs rely on enriched search output.
Models supported (see configs/models/ for the current list):
- Claude Haiku 4.5, Claude Sonnet 4, Claude Opus 4
- MiniMax M1 2.5, MiniMax M2.5
- GLM 4.7, GLM 5
- DeepSeek
- Any model supported by litellm (add a YAML config)
Prerequisites
- Python 3.11+
- Docker (for SWE-bench containers)
- Node.js 22+ (for GitNexus)
- API keys for your chosen models
Setup
cd eval
# Install dependencies
pip install -e .
# Set up API keys — copy the template and fill in your keys
cp .env.example .env
# Then edit .env and paste your key(s)
All models are routed through OpenRouter by default, so a single OPENROUTER_API_KEY is all you need. To use provider APIs directly (Anthropic, ZhipuAI, etc.), edit the model YAML in configs/models/ and set the corresponding key in .env.
# Pull SWE-bench Docker images (pulled on-demand, but you can pre-pull)
docker pull swebench/sweb.eval.x86_64.django_1776_django-16527:latest
Debug logging
Set GITNEXUS_EVAL_DEBUG=1 to include full Python tracebacks in run summaries and logs. By default, errors are sanitized to avoid leaking host paths or stack traces.
Quick Start
Debug a single instance
# Fastest way to verify everything works
python run_eval.py debug -m claude-haiku -i django__django-16527 --subset lite
Run a single configuration
# 5 instances, Claude Sonnet, native_augment mode (default)
python run_eval.py single -m claude-sonnet --subset lite --slice 0:5
# Baseline comparison (no GitNexus)
python run_eval.py single -m claude-sonnet --mode baseline --subset lite --slice 0:5
# Full Lite benchmark, 4 parallel workers
python run_eval.py single -m claude-sonnet --subset lite -w 4
Run the full matrix
# All models x all modes
python run_eval.py matrix --subset lite -w 4
# Key comparison: baseline vs native_augment
python run_eval.py matrix -m claude-sonnet -m claude-haiku --modes baseline --modes native_augment --subset lite --slice 0:50
Analyze results
# Summary table
python -m analysis.analyze_results results/
# Compare modes for a specific model
python -m analysis.analyze_results compare-modes results/ -m claude-sonnet
# GitNexus tool usage analysis
python -m analysis.analyze_results gitnexus-usage results/
# Export as CSV for further analysis
python -m analysis.analyze_results summary results/ --format csv > results.csv
# Run official SWE-bench test evaluation
python -m analysis.analyze_results summary results/ --swebench-eval
List available configurations
python run_eval.py list-configs
Architecture
eval/
run_eval.py # Main entry point (single, matrix, debug commands)
agents/
gitnexus_agent.py # GitNexusAgent: extends DefaultAgent with augmentation + metrics
environments/
gitnexus_docker.py # Docker env with GitNexus + eval-server + standalone tool scripts
bridge/
gitnexus_tools.sh # Bash wrappers (legacy — now standalone scripts are installed directly)
mcp_bridge.py # Legacy MCP bridge (kept for reference)
prompts/
system_baseline.jinja # System: persona + format rules
instance_baseline.jinja # Instance: task + workflow
system_native.jinja # System: + GitNexus tool reference
instance_native.jinja # Instance: + GitNexus debugging workflow
system_native_augment.jinja # System: + GitNexus tools + grep enrichment docs
instance_native_augment.jinja # Instance: + GitNexus workflow + risk assessment
configs/
models/ # Per-model YAML configs
modes/ # Per-mode YAML configs (baseline, native, native_augment)
analysis/
analyze_results.py # Post-run comparative analysis
results/ # Output directory (gitignored)
How It Works
Template structure
mini-swe-agent requires two Jinja templates:
- system_template → system message: persona, format rules, tool reference (static)
- instance_template → first user message: task, workflow, rules, examples (contains
{{task}})
Each mode has a system_{mode}.jinja + instance_{mode}.jinja pair. The agent loads both automatically based on the configured mode.
Per-instance flow
- Docker container starts with SWE-bench instance (repo at specific commit)
- GitNexus setup: Node.js + gitnexus installed,
gitnexus analyzeruns (or restores from cache) - Eval-server starts:
gitnexus eval-serverdaemon (persistent HTTP server, keeps LadybugDB warm) - Standalone tool scripts installed in
/usr/local/bin/— works withsubprocess.run(no.bashrcneeded) - Agent runs with the configured model + system prompt + GitNexus tools
- Agent's patch is extracted as a git diff
- Metrics collected: cost, tokens, tool calls, GitNexus usage, augmentation stats
Tool architecture
Agent → bash command → /usr/local/bin/gitnexus-query
→ curl http://127.0.0.1:4848/tool/query (fast path: eval-server, ~100ms)
→ npx gitnexus query (fallback: cold CLI, ~5-10s)
Each tool script in /usr/local/bin/ is standalone — no sourcing, no env inheritance needed. This is critical because mini-swe-agent runs every command via subprocess.run in a fresh subshell.
Eval-server
The eval-server is a lightweight HTTP daemon that:
- Keeps LadybugDB warm in memory (no cold start per tool call)
- Returns LLM-friendly text (not raw JSON — saves tokens)
- Includes next-step hints to guide tool chaining (query → context → impact → fix)
- Auto-shuts down after idle timeout
CLI flags:
| Flag | Default | Purpose |
|---|---|---|
--port <port> |
4848 |
Port to listen on |
--host <host> |
127.0.0.1 |
Bind address — use 0.0.0.0 for cross-container access |
--idle-timeout <seconds> |
0 (disabled) |
Auto-shutdown after N seconds of inactivity |
READY signal:
When the server is ready, it writes to stdout:
# IPv4
GITNEXUS_EVAL_SERVER_READY:127.0.0.1:4848
# IPv6 (bracketed to avoid colon ambiguity)
GITNEXUS_EVAL_SERVER_READY:[::1]:4848
Parse the port as the last colon-segment (split(':').pop()) — not split(':')[1], which breaks for IPv6 and for non-loopback IPv4 hosts added in this release.
Custom port and host
run_eval.py does not expose --port or --host as CLI flags. Configure them in your mode YAML under the environment: key:
# configs/modes/native_augment.yaml (or whichever mode you're running)
environment:
eval_server_port: 4849 # change if 4848 is already in use on the host
eval_server_host: "0.0.0.0" # bind all interfaces — needed for cross-container setups
Defaults are port: 4848 and host: 127.0.0.1 (loopback only). Use 0.0.0.0 only when the agent container needs to reach the eval-server from a separate network namespace. The health probe and tool scripts connect via the configured bind host (defaulting to 127.0.0.1), which is reachable for both loopback and all-interface binds.
"localhost" is also a valid eval_server_host value. The OS resolves it at bind time — typically 127.0.0.1 on dual-stack or IPv4-only systems, and ::1 on IPv6-only systems. The exact result depends on your /etc/hosts and gai.conf. The READY signal will reflect the actual bound address (e.g. GITNEXUS_EVAL_SERVER_READY:127.0.0.1:4848 or GITNEXUS_EVAL_SERVER_READY:[::1]:4848), not the literal string localhost. Use this when you want the server to bind to whichever loopback address the OS prefers rather than forcing IPv4.
Running eval-server directly in Docker / Docker Compose:
# Bind to all interfaces so sibling containers can reach it
gitnexus eval-server --host 0.0.0.0 --port 4848
# Then probe from a sibling container via its service hostname
curl http://eval-container:4848/health
If you need a non-default port (e.g. to avoid conflicts), pass --port <port> alongside --host. The READY signal will reflect both:
GITNEXUS_EVAL_SERVER_READY:0.0.0.0:5000
Parse the port as the last colon-segment (split(':').pop()) — safe for both IPv4 and bracketed IPv6 forms.
Index caching
SWE-bench repos repeat (Django has 200+ instances at different commits). The harness caches GitNexus indexes per (repo, commit) hash in ~/.gitnexus-eval-cache/ to avoid redundant re-indexing.
Grep augmentation (native_augment mode)
When the agent runs grep or rg, the observation is post-processed: the agent class calls gitnexus-augment on the search pattern and appends [GitNexus] annotations showing callers, callees, and execution flows for matched symbols. This mirrors the Claude Code / Cursor hook integration.
Adding Models
Create a YAML file in configs/models/:
# configs/models/my-model.yaml
model:
model_name: "openrouter/provider/model-name"
cost_tracking: "ignore_errors" # if not in litellm's cost DB
model_kwargs:
max_tokens: 8192
temperature: 0
The model name follows litellm conventions.
Metrics Collected
| Metric | Description |
|---|---|
| Patch Rate | % of instances where agent produced a patch |
| Resolve Rate | % of instances where patch passes tests (requires --swebench-eval) |
| Total Cost | API cost across all instances |
| Avg Cost/Instance | Cost efficiency |
| API Calls | Number of LLM calls |
| GN Tool Calls | How many GitNexus tools the agent used |
| Augment Hits | How many grep/find results got enriched |
| Augment Hit Rate | % of search commands that got useful enrichment |