commit-ko
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied commit message drafts, creating a potential surface for indirect instructions.\n
- Ingestion points: Untrusted data enters the context through the 'commit message draft' input described in SKILL.md.\n
- Boundary markers: There are no explicit delimiters or instructions to the agent to ignore embedded commands within the input text.\n
- Capability inventory: The skill's logic is focused on text substitution. It does not perform network operations, file writes, or command execution during the processing of a user request.\n
- Sanitization: The instructions do not specify any sanitization or validation of the input beyond the application of the lexicon-based transformation rules.\n- [EXTERNAL_DOWNLOADS]: The skill mentions external maintenance scripts and reference files.\n
- Evidence: SKILL.md and references/commit-lexicon.md refer to 'python3 scripts/check_commit_lexicon_ids.py' and 'skills/humanize-korean/references/ai-tell-taxonomy.md'.\n
- These are noted as tools for verifying integrity or referencing the origin of the lexicon, rather than being required for the skill's primary runtime operation.
Audit Metadata