feature-validate

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes language-specific quality gate commands such as 'make check', 'pytest', and 'npm test' based on detected project languages.
  • [COMMAND_EXECUTION]: The skill uses local platform scripts located at '/.claude/scripts/feature-state.py' and '/.claude/scripts/learning-db.py' to update feature state and log validation results.
  • [PROMPT_INJECTION]: The skill processes implementation artifacts and command outputs which presents a risk of indirect prompt injection. (1) Ingestion points: Reads implementation artifacts from '.feature/state/implement/', repository patterns from 'CLAUDE.md', and output from shell-executed test suites. (2) Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded commands within implementation files or test logs. (3) Capability inventory: The skill has access to shell execution via Bash and state management via specialized local scripts. (4) Sanitization: There is no evidence of validation or escaping of ingested data before it is incorporated into validation reports or used to drive state transitions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 05:27 PM
Security Audit — agent-trust-hub — feature-validate