systematic-debugging
Warn
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill provides instructions to access and read the .env file (cat .env | grep RELEVANT_KEY) to verify environment configurations. While intended for debugging, .env files frequently contain sensitive credentials and project secrets.- [COMMAND_EXECUTION]: The skill utilizes several shell commands (git, pnpm, docker, uv, grep, cat) to inspect the project state, manage containers, and execute tests.- [PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, such as error messages, stack traces, and request logs, which serves as a surface for indirect prompt injection.
- Ingestion points: Phase 1 (Error messages, stack traces, and request/service logs).
- Boundary markers: Absent.
- Capability inventory: Shell execution capabilities using git, pnpm, docker, uv, and grep (SKILL.md).
- Sanitization: Absent.
Audit Metadata