Agent skill

pre-action

Mandatory decision framework - clarify goal, evaluate options, choose consciously before acting

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/pre-action

SKILL.md

Skill: pre-action

What I do

I produce a PREFLIGHT before any tool calls: clarify goal, identify constraints, plan steps, and mark which calls can run in parallel. This is the PLAN phase — execution comes after via parallel-execution.

When to use me

  • Always — produce PREFLIGHT before first tool call in any task
  • Before irreversible actions (deployment, deletion, commits)
  • When facing unclear requirements or multiple viable approaches

PREFLIGHT Schema (by role)

Orchestrators (sisyphus, hephaestus, atlas, Tech-Lead):

PREFLIGHT:
  Goal: <one sentence>
  Constraints: <limits, policies, blockers>
  Plan: <≤5 numbered steps>
  Parallel: <which steps are independent>
  Stop: <conditions to halt or escalate>

Workers (Senior-Engineer, QA-Engineer, Writer, etc.):

PREFLIGHT:
  Assumptions: <what I believe to be true>
  Plan: <≤5 numbered steps>
  Parallel: <which steps are independent>
  Risks: <what could go wrong>

Read-only (explore, Researcher, Data-Analyst):

PREFLIGHT:
  Assumptions: <what I believe to be true>
  Plan: <≤3 numbered steps>
  Parallel: <which reads/searches can batch>

After PREFLIGHT

Once PREFLIGHT is complete, use parallel-execution skill to batch all independent calls identified in the Parallel field.

Mid-chain reflection (sequential tool use)

When executing a chain of sequential tool calls where each step depends on the previous result, apply a reflection step between calls:

After each significant tool result, ask:

  • Does this result change my plan?
  • Am I still on the right path, or do I need to backtrack?
  • Do I have all information needed for the next step?

Before any irreversible action, verify:

  • What exactly will this change?
  • Is this the right target (file, record, resource)?
  • Can I undo this if wrong?

When results are unexpected, stop and reassess:

  • Why did I get this result?
  • Does my mental model need updating?
  • Should I try a different approach?

This is distinct from upfront pre-action thinking — it is reactive, triggered by new information from tool results. Most valuable in long tool chains, policy-heavy environments, and sequential decisions where mistakes compound.

Related skills

  • parallel-execution — Execute phase: batch independent calls after PREFLIGHT
  • memory-keeper — Capture decision reasoning
  • critical-thinking — Rigorous analysis for complex decisions

KB Reference

~/vaults/baphled/3. Resources/Knowledge Base/AI Development System/Skills/Core-Universal/Pre Action.md

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results