maintain-code-scans-skill
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a local utility script,
scripts/check-skill-paths.sh, to verify the existence of repository paths mentioned in documentation. This script employsgit ls-fileswith argument delimiters (--) to safely query the repository index. The instructions also involve standardgit diffandgit logcommands to identify changes in the repository structure. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process the content of other agent skill files (
SKILL.md). This represents a potential surface for indirect prompt injection. However, the risk is mitigated by the skill's specific focus on structured verification (path checking and citation matching) and the shell script's internal filtering, which ignores strings containing shell-sensitive characters like parentheses, braces, and redirection symbols within backticked spans.
Audit Metadata