interop

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection because it ingests data from external plan files and uses it to drive agent actions.
  • Ingestion points: The skill reads the content of the file specified in the $ARGUMENTS variable and extracts fields like Title, Goal, and Spec (SKILL.md, Step 1 and 2).
  • Boundary markers: No delimiters or specific instructions (e.g., 'ignore embedded commands') are used when parsing or passing the extracted data.
  • Capability inventory: The skill possesses significant capabilities, including the ability to write to the local filesystem using the Edit tool, add items to a backlog via the mcp__plugin_dh_backlog__backlog_add tool, and execute other AI agent skills via the Skill tool (SKILL.md, Step 3, 4, 6, 7).
  • Sanitization: There is no evidence of sanitization or validation of the content extracted from the plan file before it is utilized in tool calls or skill delegations.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 08:41 AM