workflow-onboard

Warn

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands git log and git diff to gather context from the repository history.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses sensitive file paths, specifically .env.local and .env.example, to extract environment variable names. While the instructions explicitly state to never print values, the access to the file itself represents a potential exposure risk.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository (README files, SQL migrations, manifests). A malicious codebase could include instructions in these files to hijack the agent's behavior. 1. Ingestion points: Reads README.md, package.json, supabase/migrations/*.sql, and other repository files. 2. Boundary markers: No explicit boundary markers or 'ignore embedded instructions' warnings are used when processing the files. 3. Capability inventory: Includes shell command execution (git) and file reading across the codebase. 4. Sanitization: No sanitization or filtering of external content is specified before processing.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 25, 2026, 05:56 AM
Security Audit — agent-trust-hub — workflow-onboard