setup-matt-pocock-skills
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to inspect the local environment and manage project workflows.
- Evidence: Calls
git remote -vand reads.git/configto identify repository hosting. - Evidence: Provides templates for executing GitHub CLI (
gh) and GitLab CLI (glab) commands for issue creation and management. - [PROMPT_INJECTION]: The skill acts as a configuration engine that modifies the agent's primary instruction files (
CLAUDE.mdorAGENTS.md), creating a surface for indirect influence on agent behavior. - Ingestion points: Reads repository files including
CONTEXT.md,docs/adr/, and existing configuration files (AGENTS.md/CLAUDE.md) from the file system. - Boundary markers: None identified; the skill assumes the content of these files is valid project documentation.
- Capability inventory: Includes shell command execution via Git/GitHub/GitLab CLIs and file-write access to core agent instruction files.
- Sanitization: Content from the ingested files is used to generate new configurations without explicit sanitization or escaping mechanisms.
- [DATA_EXFILTRATION]: The skill accesses
.git/config, which may contain sensitive repository metadata. However, this data is used strictly for local environment detection and is not transmitted to external domains outside of standard authenticated CLI tool usage.
Audit Metadata