forge-reflect
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local development tools including
gitandghto identify code changes and fetch PR details. It also runs project-defined quality gates such as linting and testing scripts as specified in Step 1 and Step 3. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from code diffs and PR descriptions which could contain instructions to influence the agent.
- Ingestion points: Reads data via
git diffandgh pr viewin Step 1. - Boundary markers: The skill does not implement explicit boundary markers or instructions to ignore embedded prompts within the diff content.
- Capability inventory: The agent has the capability to execute arbitrary shell commands (via tests), modify local files (applying fixes), and interact with the GitHub API (creating issues).
- Sanitization: No sanitization is performed on the ingested code content before it is processed by the review agents.
Audit Metadata