reconcile-spec

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes content from various local files which may contain adversarial instructions.
  • Ingestion points: The agent reads spec.md, decisions.md, seed.md, and proposal.md from the project's thread directory.
  • Boundary markers: There are no defined delimiters or instructions to treat the file content as data only, increasing the risk that the LLM might follow instructions embedded within these files.
  • Capability inventory: The skill has the ability to modify spec.md on the local filesystem and invoke the /emit-pending-decisions tool.
  • Sanitization: The instructions do not specify any validation or sanitization of the input text before it is used to update the specification file.
  • [NO_CODE]: No executable scripts, binaries, or other code files are included in the skill package; it is comprised entirely of instructions and configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 01:01 PM
Security Audit — agent-trust-hub — reconcile-spec