This release includes 1 security fix for security teams reviewing exposed deployments.
Published 1mo
Developer Productivity
✓ No known CVEs patched
This release patches 1 known CVE
Topics
ai-coding-assistant
archetype
ast-analysis
claude-code
claude-plugin
code-conventions
+12 more
code-review
code-style
developer-tools
javascript
linter
mcp
mcp-server
prompt-engineering
python
ruby
rails
typescript
Affected surfaces
auth
Summary
AI summaryFixed inconsistent arming of the Stop backstop for non-code files after reaching correction limits.
Full changelog
Fixed
- The corrections-exhausted breaker no longer arms the Stop backstop on a
non-code file's credential. Once a file is corrected
MAX_CORRECTIONS_PER_FILEtimes, advisory feedback is suppressed but a
deterministic-hard secret still arms the Stop backstop so a credential cannot
slip in unblocked. Every other arming site and both Stop re-lint branches run
that secret throughblock_eligible_on_file(..., language=detect_language()),
so a credential-shaped token in markdown / config prose (no recognized
language) stays advisory and never arms the backstop — such a file has no
inlinechameleon-ignoreescape and the re-lint drops it anyway. This breaker
site skipped that gate, so a non-code file that resolved to an archetype (via a
legacy extension-blindpaths_pattern) at the corrections cap could be armed
inconsistently with the re-lint that then clears it. It now arms only on a
recognized code language, matching the sibling sites. No behavior change for
code files.
Security Fixes
- Prevents the Stop backstop from arming on credentials in non-code files after `MAX_CORRECTIONS_PER_FILE` is reached, aligning behavior with code-file handling.
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About Chameleon
All releases →Beta — feedback welcome: [email protected]