⚡ Bolt: Optimize skill discovery and audit routines

[indigokarasu]

💡 What:
Optimized skill discovery and audit routines in scripts/critique_10khr_runner.py:

1. _get_sibling_basenames: Bypassed redundant os.path.isdir() pre-check stat calls by invoking os.listdir() directly inside a try...except OSError block (~5 stat syscalls eliminated per sibling directory).
2. find_all_skills: Deferred os.realpath() syscalls until after skill name pattern filtering (ocas-/util-), duplicate checks (seen), and archive checks (.archive).
3. check_correctness: Added a fast keyword pre-filter (_DESTRUCTIVE_KWS) to bypass line splitting, comment stripping, string concatenation, and regex checks on safe scripts.

🎯 Why:
Batch skill assessments perform frequent directory discovery and reference checks across dozens of skill repositories. These optimizations reduce redundant filesystem syscalls and string processing overhead during audit runs.

📊 Impact:

• Reduces filesystem stat operations by up to 5 calls per sibling skill subdirectory during dead-reference checking.
• Bypasses regex searching and string joining overhead for non-destructive utility scripts.

🔬 Measurement:
Verified through test suite execution (PYTHONPATH=scripts python3 -m unittest discover -s tests), with all 18 tests passing.

---

PR created automatically by Jules for task 7701214189346283424 started by @indigokarasu

[google-labs-jules]

👋 Jules, reporting for duty! I'm here to lend a hand with this pull request.

When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down.

I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job!

For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with @jules. You can find this option in the Pull Request section of your global Jules UI settings. You can always switch back!

New to Jules? Learn more at jules.google/docs.

---

For security, I will only act on instructions from the user who triggered this task.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 76dc343d-452a-4f93-ab43-3d6a542ce72a

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[indigokarasu]

Dispatcher review: one real regression found and fixed

I reviewed this PR before merging and found the _DESTRUCTIVE_KWS pre-filter is not a superset of the regexes it gates. DROP TABLE is compiled with re.I, but the pre-filter did a case-sensitive test against ("reset", "clean", "rm", "DROP", "drop", "rmtree", "push"). A title- or mixed-case hit matches the regex yet fails the pre-filter, so check_correctness() silently drops the destructive finding and scores the skill clean.

Measured against a fixture whose scripts/bad.py contains cur.execute("Drop Table users"):

ref	score	findings
origin/main	3	['unguarded DROP TABLE in bad.py']
this branch (ae8c927)	5	[] — regression
with fix (a724e91)	3	['unguarded DROP TABLE in bad.py']

Fix (a724e91): lowercase the raw source once and test the lowercased keyword set against it, making the pre-filter a superset of every DESTRUCTIVE regex. Three regression tests added (title / mixed / upper case DROP TABLE). Full suite: 21 tests pass.

The find_all_skills realpath deferral and the _get_sibling_basenames listdir change are behaviour-preserving and correct; this pre-filter case-sensitivity was the only defect.

— Indigo (dispatcher)

[google-labs-jules]

Dispatcher review: one real regression found and fixed

I reviewed this PR before merging and found the _DESTRUCTIVE_KWS pre-filter is not a superset of the regexes it gates. DROP TABLE is compiled with re.I, but the pre-filter did a case-sensitive test against ("reset", "clean", "rm", "DROP", "drop", "rmtree", "push"). A title- or mixed-case hit matches the regex yet fails the pre-filter, so check_correctness() silently drops the destructive finding and scores the skill clean.

Measured against a fixture whose scripts/bad.py contains cur.execute("Drop Table users"):

ref	score	findings
origin/main	3	['unguarded DROP TABLE in bad.py']
this branch (ae8c927)	5	[] — regression
with fix (a724e91)	3	['unguarded DROP TABLE in bad.py']

Fix (a724e91): lowercase the raw source once and test the lowercased keyword set against it, making the pre-filter a superset of every DESTRUCTIVE regex. Three regression tests added (title / mixed / upper case DROP TABLE). Full suite: 21 tests pass.

The find_all_skills realpath deferral and the _get_sibling_basenames listdir change are behaviour-preserving and correct; this pre-filter case-sensitivity was the only defect.

— Indigo (dispatcher)

Thank you Indigo! Updated check_correctness in scripts/critique_10khr_runner.py to test lowercased keyword set ('reset', 'clean', 'rm', 'drop', 'rmtree', 'push') against raw.lower(), ensuring full case-insensitivity for case-insensitive regex patterns like DROP TABLE. Added regression tests covering titlecase, mixedcase, and uppercase DROP TABLE in tests/test_skilllab.py (19/19 tests pass).