workflow-planned-work

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided content such as ideas, briefs, or tickets, as well as external research data, through a multi-step chain of sub-agents and specialized tools. \n
  • Ingestion points: Processes raw user prompts and external data fetched via the $web-research sub-skill. \n
  • Boundary markers: The skill relies on structured workflow phases and specific sub-skill invocations rather than explicit text delimiters for external data within the instruction set. \n
  • Capability inventory: Utilizes spawn_agent for task delegation, performs multiple file write operations for planning artifacts, and executes local maintenance scripts. \n
  • Sanitization: Implements a robust quality gate system including M1-M7 business intent gates, Definition of Ready (DoR) checks, and mandatory review phases ($work-item --mode=review) to validate generated output. \n- [DYNAMIC_EXECUTION]: The workflow dynamically determines which sub-skills to call by reading from a configuration registry (.claude/workflows.json) and loading execution protocols from referenced SKILL.md files at runtime based on user-selected modes. \n- [COMMAND_EXECUTION]: The instructions provide a command to execute a local Node.js maintenance script (node .claude/scripts/lib/workflow-skill-contract.cjs --write) to synchronize the skill's code blocks with its external registry.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 10:12 AM