grimoire-reporter

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it incorporates data from requirement specifications and conversations directly into its reasoning and output generation.
  • Ingestion points: Requirement sources are read from files located in .grimoire/spec/ and from the user conversation history during the decomposition process.
  • Boundary markers: There are no explicit instructions for the agent to use delimiters or to disregard instructions that might be embedded within the source requirements themselves.
  • Capability inventory: The skill utilizes file-reading capabilities for repository context and file-writing capabilities specifically for the .grimoire/ticket/ directory.
  • Sanitization: No sanitization or validation logic is defined to check the contents of specification files before they are processed to generate ticket goals and implementation approaches.
  • [SAFE]: The skill does not execute shell commands, perform network requests, or download external code. Its activities are confined to local file system operations related to project documentation.
  • [NO_CODE]: This skill consists entirely of instructional markdown and reference templates; it does not ship with or execute any external scripts or binary code.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 01:22 PM