setup-matt-pocock-skills
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git remote -vand reads.git/configto identify the repository's hosting platform (GitHub or GitLab). - [COMMAND_EXECUTION]: The skill provides instructions for interacting with the GitHub CLI (
gh) and GitLab CLI (glab) to manage issues and pull requests. These are standard, well-known tools for these services. - [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for potentially untrusted data that influences future agent behavior.
- Ingestion points: repository remote URLs, directory structures, existing agent configuration files (
CLAUDE.md,AGENTS.md), and user-provided descriptions for custom issue tracker workflows. - Boundary markers: The skill does not explicitly use delimiters when incorporating findings or user descriptions into the generated documentation files.
- Capability inventory: The skill writes configuration to
docs/agents/and modifies central agent instruction files (CLAUDE.mdorAGENTS.md). It also executes CLI tools for repository management. - Sanitization: No explicit sanitization or validation logic is applied to the metadata or user descriptions before they are written to the documentation files.
Audit Metadata