harness
Pass
Audited by Gen Agent Trust Hub on Jul 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute repository-native commands such as 'make', 'git', and 'curl' for verification and environment setup. This is a primary functional requirement for a repository auditing tool and follows local development conventions.
- [PROMPT_INJECTION]: The audit workflow involves reading untrusted repository documentation like 'AGENTS.md' and 'CLAUDE.md', which presents a potential surface for indirect prompt injection. Ingestion points: Reads instructions from repository-local documentation files including AGENTS.md, CLAUDE.md, and files in the docs/ directory (SKILL.md). Boundary markers: None identified; repository content is ingested directly into the agent context. Capability inventory: The skill authorizes shell command execution via make targets, git, and curl to verify the environment (SKILL.md). Sanitization: No specific sanitization is mentioned; the skill relies on repository-native validation and the agent's internal safety guardrails.
Audit Metadata