skills/desek/governance/implement-cr/Gen Agent Trust Hub

implement-cr

Pass

Audited by Gen Agent Trust Hub on Oct 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and implement instructions from Change Request markdown files (CRs). This creates an attack surface where a maliciously crafted CR could contain instructions intended to influence the agent's behavior, subagent creation, or code implementation.
  • Ingestion points: The skill reads files located at docs/cr/CR-*.md as the primary source of implementation logic (step 1 and 4).
  • Boundary markers: There are no instructions to use delimiters or ignore embedded natural language instructions within the CR files.
  • Capability inventory: The skill possesses significant capabilities, including creating subagents via the Workflow tool, modifying the local file system (docs and code), and performing Git commits via the /checkpoint-commit command.
  • Sanitization: No sanitization or validation of the input CR files is performed before they are parsed and implemented by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 30, 2026, 06:49 PM