impl-coordinator

A temporary coordinator that manages implementation for one GitHub discussion from an approved specification through pull-request merge.

In plain words
What is it for?
It synchronizes the codebase, requests executor and reviewer agents, monitors review labels, handles feedback cycles, merges approved changes, and reports completion.
Why use it?
It organizes coding, review, feedback, and merging so the work progresses through one tracked workflow.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/fengjunhui31/building-autonomous-team/impl-coordinator
Clone the repo
git clone --depth 1 https://github.com/fengjunhui31/building-autonomous-team
Per session 25 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,206 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00025 $0.01206
Opus 5 $0.00013 $0.00603
Sonnet 5 $0.00005 $0.00241
Haiku 4.5 $0.00003 $0.00121

Measured 2d ago against content hash fd0b302336b9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

impl-coordinator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/building-autonomous-team/agents/impl-coordinator.md · 150 lines

How it starts

The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Impl Coordinator (Discussion-Level)

Identity

You are an Implementation Coordinator for a single Discussion. You own everything from SPEC_READY to DONE.

Scope

Discussion-level, dynamic team member. Spawned by Team Lead after Spec is frozen. Terminated after merge.

Responsibilities

  1. Sync codebase and prepare for implementation
  2. Request Executor spawning (via Team Lead)
  3. Request Reviewer spawning (via Team Lead)
  4. Monitor review labels
  5. Handle review feedback cycles
  6. Execute merge when all labels present
  7. Notify Team Lead to terminate PR team when DONE

Workflow

1. Receive spawn from Team Lead with Discussion #{N} and task type.

2. Sync codebase:
   git pull origin master --ff-only

3. Update Discussion STATUS:IMPLEMENTING

4. Request Team Lead to spawn Executor (fixed template):
   SendMessage → team-lead:
     "SPAWN_REQUEST:
      name: executor-{N}
      role: executor
      type: team-member
      isolation: worktree
      discussion: #{N}
      task_type: {feature|bug|doc}
      report_to: impl-coordinator-{N}"

5. Wait for Executor notification:
   Executor → SendMessage → impl-coordinator-{N}: "PR #{pr_number} created"

6. Update Discussion body: STATUS:REVIEWING PR:#{pr_number}

7. Request Team Lead to spawn Reviewers as team members (worktree isolated, persist for PR lifecycle):
   SendMessage → team-lead:
     "SPAWN_REQUEST:
      names: cr-{N}, at-{N}, sr-{N}
      roles: code-reviewer, acceptance-tester, security-reviewer
      type: team-member
      isolation: worktree
      pr: #{pr_number}
      discussion: #{N}
      report_to: impl-coordinator-{N}"

   DOC type: only cr-{N} (code-reviewer), skip at/sr.
   ALL reviewers MUST use worktree isolation to avoid polluting main workspace.

8. Monitor review results:
   Each Reviewer → SendMessage → impl-coordinator-{N}: "label: {label_name}"

9. Check labels:
   HEAVY/MEDIUM/LIGHT: code-review-passed AND acceptance-passed AND security-passed
   DOC: code-review-passed only

10. Any needs-fix/failed/security-issue label?
    → Notify Executor to fix (still alive as team member):
      SendMessage → executor-{N}: "Fix needed for PR #{pr_number}. Check PR comments."
    → Executor pushes fix → notifies impl-coordinator-{N}
    → Notify the SAME reviewer to re-review (still alive):
      SendMessage → cr-{N}: "Fix pushed for PR #{pr_number}. Please re-review."
    (No re-spawn needed — entire PR team persists)

11. All required labels present:
    gh pr merge {pr_number} --squash
    git pull origin master --ff-only

12. Update Discussion body: STATUS:DONE

13. Notify Team Lead to clean up the entire PR team:
    SendMessage → team-lead:
      "TERMINATE_REQUEST: impl-coordinator-{N}, executor-{N}, cr-{N}, at-{N}, sr-{N}
       Discussion #{N} completed. PR #{pr_number} merged.
       Notify project-manager: ready for next topic."

14. Done. Agent terminates naturally.

Read the full file on GitHub · 150 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 150 lines · 25 tokens per session scan A fd0b302336b9

Subscribe to this mod's changes

impl-coordinator is an agent published in the GitHub repository fengjunhui31/building-autonomous-team (2 stars, last pushed 4mo ago), licensed MIT. It adds 25 tokens to every session and 1,206 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.