iam-temp-delegation-review

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches AWS Service Authorization Reference (SAR) data from https://servicereference.us-east-1.amazonaws.com/. This is an official AWS domain. The implementation in src/iam_delegation_review/sar_lib/_feed.py includes a validation utility (_validate_feed_url) that strictly allowlists the *.amazonaws.com suffix and enforces HTTPS.
  • [COMMAND_EXECUTION]: The skill uses standard AWS CLI and Python commands for environment setup and execution (e.g., pip install, aws sts get-caller-identity). These are necessary for its documented purpose and target a controlled environment (virtualenv).
  • [DATA_EXFILTRATION]: No evidence of unauthorized data exfiltration. Network activity is limited to official AWS endpoints (Access Analyzer and the SAR feed). Local state is managed within a registry/ directory with path validation (_validate_path_component) to prevent directory traversal.
  • [PROMPT_INJECTION]: The skill uses detailed instructional documents (procedure-reviewer.md and procedure-verifier.md) to guide the agent. It emphasizes grounding analysis in authoritative sar_context.json data, which mitigates risks associated with processing potentially untrusted IAM policy documents.
  • [SAFE]: The code follows professional software engineering practices, including schema validation for its configuration files, use of locks for concurrent network access, and a clear multi-stage pipeline with deterministic gates.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 10:54 PM
Security Audit — agent-trust-hub — iam-temp-delegation-review