⚡AgentSkills
🔄 Agentic Methodology · Execution Discipline

Dispatch subagents with fresh-context discipline

Fresh implementer per task, brief files as single source of truth, two-stage review, capped fix loops.

advanced~40 minAgent Power UsersAI-Native EngineersTeams Shipping With Agents

Steps

  1. 1Extract each task's brief to its own file; the brief is the ONLY requirements source
  2. 2Dispatch a FRESH subagent per task — context pollution breeds confident errors
  3. 3Require two-stage review per task: spec compliance first, then code quality
  4. 4Cap fix loops at three resumes; beyond that, fresh agent + more capable model
  5. 5Batch trivially-same-shaped small edits into ONE dispatch, reviewed as one diff
  6. 6Keep a findings ledger: parked minors get adjudicated, never silently dropped

Common Pitfalls

  • ▲Session-model inheritance silently spending premium tokens on mechanical tasks
  • ▲Fix loops where the same blind implementer retries identical failures

Commands

Install with skills CLI
$ npx skills add aniruddhaadak80/skills --skill execution-discipline-subagent-dispatch-discipline
Install globally
$ npx skills add aniruddhaadak80/skills --skill execution-discipline-subagent-dispatch-discipline -g

Tags

#subagents#orchestration#quality#agentic-methodology#execution-discipline

Related skills

Failing test first, watch it fail, minimal code, watch it pass, commit — no code born before tests.

🔄 Agentic Methodology·~30m

Fresh-command proof for every completion claim; 'it should work' is a confession.

🔄 Agentic Methodology·~20m