fix(monitor): extend high-FP set after auditing the residual verdicts (#30)#35
Merged
Merged
Conversation
…#30) Completed the residual single-rule MALICIOUS audit. Ran each remaining rule against benign inputs; six more over-flag legitimate content and now get the verify-or-downgrade treatment: - MALWAR-CMD-002 3 legit 'npx -y @mcp/...' setup lines -> MALICIOUS - MALWAR-CRED-002 onboarding prose ('enter your API key', 'paste your token') - MALWAR-EXFIL-001 skill-tooling reads of ~/.claude/ and .cursor/ - MALWAR-EXFIL-003 'curl -X POST -d "$(cat file)"' (normal API upload) - MALWAR-PI-001 security/detection skills that quote injection phrases - MALWAR-HIJACK-001 roleplay/persona skills ('you are now...', 'your new role is') Held up and stay confident on a single hit (no FP in testing): PERSIST-001 (cron/systemd/.bashrc), EXFIL-002 (crypto-wallet), FRAUD-002 (front-running). Recall preserved: all still fire on the malicious fixtures; corroborated (2+ rule) verdicts unaffected. On the current snapshot this leaves only 20 unverified single-rule MALICIOUS (PERSIST-001 x15, EXFIL-002 x4, FRAUD-002 x1), all on rules that did not over-flag. Full unit suite (1445) passes. Co-Authored-By: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Completes the residual single-rule MALICIOUS audit from #30. Ran each remaining
rule against benign inputs; six more over-flag legitimate content and now
get the verify-or-downgrade treatment (join
HIGH_FP_RULES):CMD-002npx -y @mcp/...setup lines → MALICIOUSCRED-002EXFIL-001~/.claude/and.cursor/EXFIL-003curl -X POST -d "$(cat file)"(normal API upload)PI-001HIJACK-001Held up and stay confident on a single hit (no FP in testing):
PERSIST-001(cron/systemd/.bashrc),EXFIL-002(crypto-wallet access),FRAUD-002(agentic front-running).Impact
(2+ rule) verdicts unaffected.
39 → 20 (now only PERSIST-001 ×15, EXFIL-002 ×4, FRAUD-002 ×1 — all on
rules that did not over-flag).
Closes the audit item of #30
The remaining 20 sit on rules with no demonstrated FP surface. Other #30 items
(detector novelty logic, human-approval gate, dry-run mode) are still open.
🤖 Generated with Claude Code
Generated by Claude Code