maintain-verification-skill
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXPOSURE]: The skill searches for files within the user's home directory at specific paths (~/.agents/skills/ and ~/.claude/skills/). While these paths are related to agent configurations, accessing the home folder represents a sensitive capability for locating and reading local data.
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves reading and processing external source code and other skill definitions, which constitutes an attack surface for indirect prompt injection.
- Ingestion points: The agent reads target skill files and project source code during the discovery and source wave phases (SKILL.md Steps 0 and 2).
- Boundary markers: The instructions lack explicit delimiters or specific warnings to ignore instructions that might be embedded within the project files being audited.
- Capability inventory: The skill possesses the ability to spawn subagents, write files to the local directory, and perform network-facing operations via the GitHub CLI.
- Sanitization: No sanitization or validation of the ingested content is performed before it is processed by the agent or used to generate summaries and PR content.
- [COMMAND_EXECUTION]: The skill invokes external command-line tools, specifically using
gh pr createto submit pull requests to remote repositories.
Audit Metadata