sem-inspect-weave

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes untrusted content from external repositories. Ingestion points: Pull request content and repository data are ingested via inspect pr and inspect grep as defined in references/commands.md and scripts/inspect-review-target. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the wrapper scripts. Capability inventory: The skill provides access to local binaries sem, inspect, weave, and gh through the scripts/ directory. Sanitization: External output from the tools is not sanitized or validated before being returned to the agent context.
  • [COMMAND_EXECUTION]: The skill uses bash scripts to wrap and execute local semantic analysis tools. These commands are restricted to the intended functionality of the sem, inspect, and weave binaries.
  • [SAFE]: The documentation and scripts include explicit safety warnings against running commands that mutate repository configuration, such as weave setup, unless specifically requested by the user.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 08:41 PM