Agent finish line
Skill mrcha033/skills/plugins/agent-finish-line/skills/agent-finish-line
Finish an implementation, experiment, migration, release, or debugging task with the least process needed. Use when work is expanding, repeating unchanged attempts, or stopping before the requested deliverable; drive it to completion or one evidence-backed external blocker without creating task-management artifacts.From its SKILL.md
npx -y skills add mrcha033/skills --skill agent-finish-lineAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
2.3 KB, as published. Nobody here has run it
Agent Finish Line
Drive the user's requested endpoint to completion. This skill changes execution discipline only; it does not introduce a contract format, receipt, ledger, or separate workflow.
Work
- Identify the requested endpoint and the earliest unmet condition internally.
- Perform the smallest action that materially advances or decides that condition.
- Verify the result in proportion to its risk.
- If the same observable failure repeats, change strategy instead of retrying it unchanged.
- Route optional improvements out of scope and stop when the requested endpoint is delivered.
Preserve any terminal condition the user supplied. Keep safety, authorization, and validation rules from the task's domain; this skill does not weaken them.
No ceremony
- Do not create
.agent-finish*files or any contract, gate, receipt, backlog, state-signature, or attempt-tracking artifact merely because this skill was invoked. - Do not ask the user to sign, approve, or acknowledge an internal execution plan.
- Do not announce internal gates or checklists unless doing so helps the user or the user explicitly asks for them.
- Do not add a second audit layer after decisive verification passes.
- Reuse task-native evidence such as a test result, commit, deployed endpoint, or generated output. Do not manufacture evidence paperwork around it.
- Create a formal acceptance record only when the user explicitly requests one or the target system itself requires it.
Finish
Complete when the requested deliverable exists and the decisive verification passes. If completion depends solely on unavailable external authority, state, credentials, hardware, or a required user choice, report:
- the unmet endpoint;
- the observed evidence;
- the exact missing external condition;
- the smallest action that unblocks it.
Lead the final response with what finished or what remains blocked. Keep it short unless the user asks for detail.
What ships with it: 1 file
222 B alongside SKILL.md
agents/
- openai.yaml222 B
Gives 0 of the 12 instructions most context ai engineering skills give
Counted across 1,328 of the 2,349 authors here whose files we hold, read 2026-09-06
- Dispatch a fresh subagent for each taskin 76 of 1328, across 59 files
- Perform spec compliance review before code quality reviewin 44 of 1328, across 34 files
- Dispatch a final code reviewer after all tasksin 38 of 1328, across 26 files
- Answer subagent questions before allowing implementationin 36 of 1328, across 26 files
- Use the least powerful model capable of the taskin 33 of 1328, across 26 files
- Create a TodoWrite list for all tasksin 32 of 1328, across 22 files
- Perform a task review after each implementationin 31 of 1328, across 24 files
- Extract all tasks and context from the planin 29 of 1328, across 20 files
- Provide full task text to subagentsin 28 of 1328, across 20 files
- Use git worktrees for isolated workspacesin 25 of 1328, across 20 files
- Specify the model explicitly when dispatching a subagentin 23 of 1328, across 18 files
- Execute all tasks from the plan without stoppingin 21 of 1328, across 16 files
Said here and by no other author read
- Identify the earliest unmet condition
- Perform the smallest action to advance the condition
- Change strategy if failures repeat
- Route optional improvements out of scope
- Stop when the requested endpoint is delivered
- Preserve domain safety and validation rules
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once.