ns-harness-codebase-reverse-spec
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external codebase content, creating a surface for indirect prompt injection attacks. 1. Ingestion points: The skill uses tools to read files and list directories as defined in SKILL.md. 2. Boundary markers: Absent; there are no instructions to use specific delimiters or ignore-instructions warnings when processing code content. 3. Capability inventory: The agent possesses file system access and can execute the local script scripts/scan_leakage.sh as instructed in SKILL.md. 4. Sanitization: Absent; no validation or escaping of codebase content is performed before processing.
- [COMMAND_EXECUTION]: The skill instructs the agent to run an included shell script to scan generated documentation for technical leakage. 1. Evidence: Phase 4 of the workflow in SKILL.md suggests running scripts/scan_leakage.sh. 2. Analysis: The script is a static bash utility provided within the skill package that performs simple regex matches; it does not execute dynamic or remote code.
Audit Metadata