improve-environment
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from local agent logs and project-level configuration files to identify friction points and propose changes. 1. Ingestion points: Identified in SKILL.md Step 1 (agent logs) and Step 2 (AGENTS.md, conventions, package.json). 2. Boundary markers: The instructions do not define clear boundaries or 'ignore' instructions for the data being read. 3. Capability inventory: The skill can execute arbitrary project scripts (lint, test) and modify local repository configuration (core.hooksPath) in SKILL.md Step 5. 4. Sanitization: There is no mention of sanitizing or escaping the content read from logs or project files before it influences the agent's actions.
- [COMMAND_EXECUTION]: The skill explicitly directs the agent to run project-defined scripts for linting, type-checking, and testing during the environment audit and verification phases in SKILL.md.
- [PERSISTENCE]: The skill establishes automated checks by configuring local git hooks using the core.hooksPath setting in SKILL.md Step 5, which is an intended feature to provide project guardrails.
Audit Metadata