[irvingpop]

Fixes the false "over the download byte budget" blocks on front-end#2027 (42 packages at release cf19216), and adds the local check the owner asked for before any dependency release.

Cause (reproduced locally). bounded_fetch reserved the full 128 MiB per-response cap before each download. Eight workers held the whole 1 GiB budget while only about 390 MB had arrived; a refused reservation failed instantly, so one worker drained the rest of the queue in about 25 ms. A race, not an overspend: the PR needs 577.6 MB decoded (71.5 MB gzip), and lookups took 9 s of 240 s. Local harness at cf19216: 3, 0 and 2 false failures in three live runs (timing-dependent); a deterministic unit test fails 28 of 40 keys on the old code.

Fix.

• No reservation: count decoded bytes actually received (including partial reads from failed attempts, once per attempt, under the lock); refuse a new fetch only when received >= budget. Overshoot bounded by workers x per-response cap.
• MAX_REGISTRY_BYTES = 8 GiB decoded (about 14x this PR); the 240 s run budget and the per-response cap still bound time and memory.
• Kept registry data per version holds at most 1,000 characters of each lifecycle script, visibly marked as cut for the model. Detection of new or changed install scripts uses script names and the full package.json text, not the cut text.
• Design doc K-20, K-21; K-18/K-19 corrected.

New tool. review/tools/dep_repro.py runs the same collect dependency stage (same parser, delta, limits, worker count) on two lockfile blobs fetched through the REST API, against live registries. It never runs PR code. Use it before any dependency-check release.

Local validation. front-end#2027: 0 hard blocks in 6 harness runs (206/206 packages). Controls (back-end, operationcode-pybot, claude-review-test, front-end, a one-package change): no lookup failures; only true 72-hour fresh-publish blocks. Stress: real budget overrun, oversized document, gzip bomb and 8 large concurrent downloads behave as designed. Unit tests 2711 passed; release tests 2 passed. Eval not run.

Review. Adversarial review (other model family) found 2 Important + 1 Minor; all fixed with regression tests that fail on the previous commit; scoped re-review clean, including a trace of every use of the cut script text. Open minor: the cut marker is plain text inside the script string and the prompts do not explain it (it cannot lower risk).

Goes live only with a release.

🤖 Generated with Claude Code

[github-actions]

🔴 FAIL: OC Claude review did not complete

Caution

Incomplete file coverage blocks this PR. No waiver can clear it.
Next step for the author: fix the cause named in the row, then push. The review runs again on the new commit.

Severity	Effect	Where	Finding	Found by
🚫 Hard block	Blocks, no waiver	review/tests/unit/test_collect.py	Incomplete file coverage: file content not scanned: unterminated-private-key, in review/tests/unit/test_collect.py	File checks (code, no model)
⚠️ Not run	Blocks, no waiver	Review pass	authentication failed. The pass returned no result.	Workflow (code, no model)
⚠️ Not run	Blocks, no waiver	Security pass	authentication failed. The pass returned no result.	Workflow (code, no model)
⚠️ Not run	Blocks, no waiver	Supply chain pass	authentication failed. The pass returned no result.	Workflow (code, no model)
⚠️ Not run	Blocks, no waiver	Gate	A required job did not succeed: dependency-review: failure	Workflow (code, no model)
⚠️ Not run	Blocks, no waiver	Gate	Dependency review did not succeed: job result failure	Workflow (code, no model)
⚠️ Not run	Blocks, no waiver	Gate	Dependency review did not succeed: step outcome failure	Workflow (code, no model)

File checks	Review	Security	Supply chain	Dependency review	Coverage
🚫 1 hard block, 1 signal	❌ failed	❌ failed	❌ failed	❌ failed	🚫 incomplete

_{Run 40988632863, attempt 1 for head f11bea8 into main. The required workflow in the merge box is the result that counts. This bot never approves a PR.}

📋 Full review record: reasons, checks, signals, finding details, model summaries, token usage

Reviewed snapshot

Head SHA	Merge base SHA	Base branch	Run
f11bea8a0204237b47e0debaa83506af93f753c6	cf192165e1306772370f64af96b0d6d2e53da0af	main	40988632863, attempt 1

Reasons the gate failed

Reason	Waivable?	Detail
A required job did not succeed	🚫 no	dependency-review: failure
Dependency review did not succeed	🚫 no	job result failure
Dependency review did not succeed	🚫 no	step outcome failure
A review pass failed	🚫 no	pass review failed: authentication failed
A review pass failed	🚫 no	pass security failed: authentication failed
A review pass failed	🚫 no	pass supply-chain failed: authentication failed
Hard block from the file checks	🚫 no	incomplete-coverage in review/tests/unit/test_collect.py
Hard block from the file checks	🚫 no	incomplete-coverage (coverage record is not complete)

No waiver line is shown. At least one reason cannot be waived, so a waiver would not change the result.

Checks (code, no model)

Check	Result	Detail
Hard block	🚫 blocks	Incomplete file coverage: file content not scanned: unterminated-private-key, in review/tests/unit/test_collect.py
Coverage	🚫 NOT complete	files in comparison 6, files collected 6, trees complete
Unavailable file	⚠️ unavailable	review/tests/unit/test_collect.py: unterminated-private-key
Dependency changes	✅ none
Dependency review job	❌ failed

Signals (context only, never block)

• ℹ️ Very long line: lines longer than 500 characters, in docs/claude-review-design.md, lines 287 294 718 996 997 998 999

Dependency changes

None.

Finding details

Model output from untrusted input. Check it before you act on it. Verified means only that the quoted text is at that place. It does not mean the claim is true.

Blocking is decided by code from category and severity, never by the model. A finding in a blocking category blocks only at high or critical severity.

No findings from the passes that ran. The Review, Security, Supply chain pass did not run, so this is not a clean result.

Model summaries

Model output from untrusted input. Check it before you act on it. Verified means only that the quoted text is at that place. It does not mean the claim is true.

Pass	Status	Model	Summary
Review	❌ failed		No summary. Reason: authentication failed
Security	❌ failed		No summary. Reason: authentication failed
Supply chain	❌ failed		No summary. Reason: authentication failed

Token usage

Pass	Input	Output	Cache write	Cache read	Attempts
Review	not recorded	not recorded	not recorded	not recorded	not recorded
Security	not recorded	not recorded	not recorded	not recorded	not recorded
Supply chain	not recorded	not recorded	not recorded	not recorded	not recorded
Recorded total	not recorded	not recorded	not recorded	not recorded	not recorded