Skip to content

fix: correct over-broad languages: frontmatter in 12 core rules - #126

Merged
thomas-bartlett merged 2 commits into
cosai-oasis:mainfrom
ibrahimssmalik:fix/rule-language-targeting
Sep 29, 2026
Merged

thomas-bartlett merged 2 commits into
cosai-oasis:mainfrom
ibrahimssmalik:fix/rule-language-targeting

Conversation

@ibrahimssmalik

@ibrahimssmalik ibrahimssmalik commented Sep 3, 2026 •

Copy link
Copy Markdown
Contributor

Closes #121 - See docs/language-targeting-audit.md for per-rule before/after and rationale.

…i-oasis#121)

Removed 21 language entries that had no grounding in rule body content —
no code fences, no ecosystem references, no language-specific guidance.
Before: JavaScript triggered 18/20 rules (~21,000 tokens) for any JS repo.
After: JavaScript triggers 16 rules; spurious entries (MATLAB, Perl, Vlang,
C in web rules, YAML in code rules) are gone.

Methodology: read each contested rule body and removed only entries that
appeared nowhere in the prose, code fences, or technology references.
No additions were made. Generated skills/ and SKILL.md via convert_to_ide_formats.py.

Adds docs/language-targeting-audit.md with before/after table and per-rule
rationale for each of the 12 modified rules.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LSQzfxR8wq3uhrUtCkNTh3
@thomas-bartlett

Copy link
Copy Markdown
Contributor

Thanks for the careful audit! I pushed a small follow-up commit that removes c from the framework rule and drops the audit doc. Everything else is as you had it.

@thomas-bartlett
thomas-bartlett merged commit 5c3bcae into cosai-oasis:main Sep 29, 2026
2 checks passed
@thomas-bartlett thomas-bartlett mentioned this pull request Sep 29, 2026
6 of 7 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Rule languages: lists over-target mainstream languages, inflating skill context cost

2 participants