AGENT SKILL

executing-plans

Executes a plan that the response model saved under writing-plans, with the response model doing every task itself, when the human user chose inline execution.

License 许可
Apache-2.0
所属仓库
AidALL/ghost-alice
最近更新
2026-11-15
Stars
★ 15
Forks
4

01 前言 Front Matter

name
executing-plans
calls
[ "hard:finishing-a-development-branch", "hard:using-git-worktrees" ]
description
Executes a plan that the response model saved under writing-plans, with the response model doing every task itself, when the human user chose inline execution.
compatibility
[ "Python 3.11+ standard library" ]

03 SKILL.md 正文

Executing Plans

The response model executes a plan that it saved under writing-plans inline, updates tasks when evidence changes the implementation, and completes the integration workflow that the human user authorized. Use subagent-driven-development when the human user authorized that execution choice and it is appropriate; tool availability alone does not replace the workflow that the human user chose.

Procedure

  1. Read the plan and its requirements. Identify missing prerequisites, contradictions and human-owned decisions before starting dependent tasks.
  2. Use the checkout that the human user authorized and apply using-git-worktrees only when isolation is needed. Do not implement on main/master without the human user's explicit consent.
  3. Create or update the host's task tracker from the plan. For each ready task, mark it in progress, execute its steps and required skills, obtain the specified evidence, then mark only the supported task complete.
  4. When an observation contradicts the plan's order, commands, behavior or file mapping, edit the affected task and update the tracker before continuing. Keep the plan as current execution instructions, without appended attempt histories. A change to the human user's objective or a boundary the human user explicitly set requires the human user's decision; an implementation correction inside the boundary does not.
  5. After all required tasks and checks satisfy the plan, invoke finishing-a-development-branch.

Failure modes and continuation

Use systematic-debugging to investigate a failure; use boundary-contract for an unresolved authority decision.

Do not execute example commands as though they were the actual plan, add unrelated refactors, or bypass a required check. Preserve valid evidence through verification-before-completion.

Output and integration

Report the implemented tasks, evidence for their acceptance criteria, changed plan decisions and unresolved dependencies. Verification-before-completion defines the completion and partial completion blocks and claim scope.

04 同仓库更多 Skills

AidALL
AidALL/ghost-alice

skill-evolution

Recommends skill changes from recorded tool-call patterns and conduct corrections asserted by the human user or observed by the response model, when the human user asks for such recommendations or their pending backlog, or authorizes that analysis.

★ 15
AidALL
AidALL/ghost-alice

session-intent-analyzer

Records the response model's new or changed interpretation of the input the response model is answering as its `interpreted-as` record before the first tool call of work, and reads that record when it can change the objective, constraints, work choice or whether earlier work continues.

★ 15
AidALL
AidALL/ghost-alice

necessity-gate

Judges whether proposed work is justified, needs a narrower scope or is unjustified, when the response model proposes new work or the human user asks whether to do it.

★ 15
AidALL
AidALL/ghost-alice

merge-companion

Reviews the installer's backups of external changes to installed skills and merges each into the new skill version, discards or defers it as the human user chooses, before a task changes or reinstalls an installed skill, or when the human user asks for that review.

★ 15
AidALL
AidALL/ghost-alice

compact-handoff

Writes a handoff of the task's objective, scope, evidence and next action when a long conversation, compaction or handover to another session or agent could lose them, and resumes the task from it afterwards.

★ 15
AidALL
AidALL/ghost-alice

writing-skills

Writes a new skill or changes an existing one in its folder (SKILL.md, references and scripts) and tests it.

★ 15