focused-fix
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructs the agent to access sensitive configuration data, specifically mentioning the
process.envobject and.envfiles. This is intended for verifying that required secrets and database strings are correctly configured during a debugging session.\n- [INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect prompt injection because it is designed to ingest and analyze untrusted data from the local development environment.\n - Ingestion points: Reads project source code, log files, test failure outputs, Sentry-style error reports, and git history (
git log).\n - Boundary markers: No explicit boundary markers or "ignore instructions" delimiters are defined for the ingested data.\n
- Capability inventory: The skill has the capability to execute shell commands (running tests, git logs) and write modifications to files (Phase 4: FIX).\n
- Sanitization: No filtering or sanitization of ingested content is specified.
Audit Metadata