agentleFS
syangasyanga/astack/skills/full-send/SKILL.md

full-send

Skill0 starsChanged 53 days agoMITView on GitHub

Take a settled task through verified PRs to merge and closeout without further input. Say "no merge" to stop at merge-ready.

14 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use full-stop, leaving-pr-comments and 6 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Nexus-Organization-InfosecSkill

    full-stop

    Shut down the Full Control phone-bridge completely — end Claude's watch loop, stop the bridge server (server.py started via ensure_server.ps1), and mark Claude offline in the app. Use whenever the user invokes /full-stop or says "stop full control", "stop the watch loop", "shut down the bridge/server", or "stop the ps1". Counterpart of [[full-control]], which starts everything.

    Also in 11 projects that use this

  • scottwaterSkill

    leaving-pr-comments

    Format comments an agent posts on GitHub on the user's behalf. Use whenever leaving a PR comment, review comment, or reply.

    Also in 8 projects that use this

  • scottwaterSkill

    sol-loop

    Review a change with GPT Sol in a loop until both agents agree. Use when the user says sol loop, or when another skill asks for a Sol review.

    Also in 8 projects that use this

  • assisjpSkill

    file-pr

    Prepare and open or update an authorized pull request with a concise problem-first description and verifiable evidence. Use when the user asks to create, file, open, or update a PR. Do not use merely to review a diff or when publication is not authorized.

    Also in 8 projects that use this

  • assisjpSkill

    takeover

    Take over an existing branch, pull request, worktree, or agent handoff while preserving the original goal, contributor credit, and concurrent work. Use when the user asks to take over, finish, rescue, or continue someone else's implementation. Do not use to replace a clear request to start new work.

    Also in 3 projects that use this

  • owidSkill

    babysit-pr

    Spawn a background agent that babysits an open PR — triggers a Codex review, watches CI, judges and fixes valid findings, replies to and resolves review threads, and loops to a cap. Never merges. Trigger when the user asks to "babysit the PR", "ask for codex review(s) and watch CI", or after pushing a substantial chunk to a PR branch.

    Also in 8 projects that use this

  • +1 more from scottwater

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Create assumed-decisions.md in the task's store, .scratch/<task>/ per Locate the store.

    line 14
  2. Distinctive:

    When multi-phase-plan step 1 would skip the plan, follow implement, then ship-pr.

    line 18
  3. Distinctive:

    Before merging, or before reporting merge-ready under "no merge":

    line 52
  4. Distinctive:

    When the default branch runs CI after merges, wait for the run on each merge commit as a background job.

    line 62
  5. Distinctive:

    Follow worktree-cleanup. If it stops with work remaining, or a gate is still open, send a push notification naming what remains.

    line 63

“Common” is measured: kept by at least 6 in 10 of the 14 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Full send

You own the task from prompt to landed PRs. The user is not watching. Invoking this skill authorizes execution, PR delivery, merging, an admin merge, follow-up issues, and lane cleanup, and delegates every pick a later PR can undo. It also lifts the ask-first step from any test method that verification needs, such as running the app, browser automation, or screenshots. Gate actions and harness approval prompts still stop. With "no merge" in the request, stop each PR at merge-ready. When a linked skill ends with its report, or stops for one of these authorizations, continue to the next step. Every other stop in a linked skill still holds.

1. Start

  1. For a simple task, or one already settled, for example by grilling, skip the questions. Otherwise ask one round of questions, each with a recommended answer. Cover open questions that block the work. Also cover production deploys, live systems, destructive actions, and secrets, when the task touches one and the request leaves it open. After this step, reach the user only through gates.
  2. Create assumed-decisions.md in the task's store, .scratch/<task>/ per Locate the store. Start it with the request and whether merging is authorized. Post its absolute path in the thread.

2. Route

  • When multi-phase-plan step 1 would skip the plan, follow implement, then ship-pr.
  • Otherwise follow multi-phase-plan with the plan in this store, then orchestrate. Record this skill's authorizations, with the full admin merge rule, and the start-round answers in the plan and standing orders. Mark each plan gate that an assumed decision settles as decided by the coordinator. Record a repository-required approval as an admin merge case, not a review gate.

3. Assumed decisions

An assumed decision is a delegated pick, such as an architect candidate, a scope call, a prototype UI variant, or a choice a reviewer or bot asks for. Take the recommended option and append it to assumed-decisions.md. On the orchestrate route, workers report their picks and the coordinator appends them.

## D<n>. <question>
- Options: <A>; <B>
- Picked: <option>, because <reason>
- Undo: <what reversing it takes, with the PR or commit>

In each PR body, list that PR's entries under "Assumed decisions", one line each.

When the user sends a message that steers the work, apply it at the next checkpoint on the orchestrate route. On the implement route, apply it before the next commit. When it overrides an assumed decision, update the entry. When the affected work has merged, deliver the change as a new PR under this skill.

4. Gates

Each of these conditions is a gate:

  • a deploy, live-system, destructive, or secrets action the start round left open
  • missing access
  • a change you cannot verify on the matching surface
  • a human change request
  • a missing approval that the admin merge does not clear
  • in-scope review defects after three fix-and-review passes
  • a review bot silent through both timers in Merge step 2

On the orchestrate route, park it through orchestrate's escalation. Otherwise write its decision request to the store's open-decisions.md, and record the answer there when it arrives. Continue unblocked work. When nothing can move, ask with the question tool if one exists, otherwise end the turn with the request. Send a push notification if the harness has one.

5. Merge

Before merging, or before reporting merge-ready under "no merge":

  1. Get a clean review-pr fix-and-review pass on the final head. A commit after the last pass needs a new pass. Treat a consequential PR as a hardening request for harden-pr.
  2. Wait for a review of the final head from each review bot seen on this PR or recent merged PRs. In T3 Code, call watch_pull_request once per PR and end the turn while checks are pending. When checks pass and a bot has not yet reviewed the final head, start a 30-minute timer as a background job (run-job or codex-background-jobs). Check pr.py status when the timer fires or a review arrives. When the timer fires first, request a re-review where the bot supports one and start a second timer.
  3. File each actionable follow-up, including followups.md entries, as a GitHub issue linked from the PR. List other notes under Deferred follow-ups in the PR body.

Admin merge. When the base branch's protection and rulesets show repository-required approvals as the only unmet requirement, and every check on the head is green, follow Finish or merge without its approval check and add --admin to its merge command. A required check that has not run is unmet. A merge queue or a human change request rules out the admin merge. Leave human reviews for humans to dismiss.

6. Close

  1. When the default branch runs CI after merges, wait for the run on each merge commit as a background job. If the PR broke it, ship a fix or a revert under this skill. On the orchestrate route, orchestrate's retro watcher does the waiting.
  2. Follow worktree-cleanup. If it stops with work remaining, or a gate is still open, send a push notification naming what remains.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.