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.
npx skills add nikolamin/claude-coordinator-kit --skill comms-registergit clone --depth 1 https://github.com/nikolamin/claude-coordinator-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/nikolamin/claude-coordinator-kit/comms-register)<a href="https://agentmods.dev/skills/nikolamin/claude-coordinator-kit/comms-register"><img src="https://agentmods.dev/badge/skills/nikolamin/claude-coordinator-kit/comms-register/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/nikolamin/claude-coordinator-kit/comms-register"><img src="https://agentmods.dev/badge/skills/nikolamin/claude-coordinator-kit/comms-register.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00047 | $0.01210 |
| Opus 5.5 | $0.00019 | $0.00484 |
| Sonnet 5.5 | $0.00009 | $0.00242 |
| Haiku 4.5 | $0.00005 | $0.00121 |
Grade A, and why
comms-register 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 28d 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.
How it starts
The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Comms register
Read the notification policy in the operating profile. For new projects, default to meaningful completion plus required decisions/blockers; honor decisions-only or requested checkpoints when specified. Quiet external notifications do not override the host's active-chat progress rules. Do not create daily briefs or send routine "starting/still working" pings unless requested. Internal progress belongs in the canonical task/event record.
Quiet reporting never means leaving a prepared decision unsent or free capacity unused. Follow the required coordinator cycle after replying: ask the next valid pending question in the same turn and take eligible authorized work. Do not manufacture a new approval request for routine work already authorized.
A completion means the agreed user outcome is usable in its stated environment, not simply that an agent finished or one repository landed. Batch related completed work into one message. If the policy is decisions-only, write completion to the record without an unsolicited ping. An urgent actionable blocker can interrupt; a repeated unchanged blocker should not.
When asked for status, answer the goal/thread the founder is discussing: goal → current result → what remains → whether the founder is needed. Use their product language; put hashes, hosts, agent counts and technical proof in the record unless needed for the question. Keep phone notifications concise and contextual. Don't turn an answer into a new approval request.
Use the bundled status view for a compact operational snapshot: running work, work awaiting
release, the presented question, queued-question count and the next pending tasks. Running tasks
never appear as "next"; user-initiated actions do not inflate "waiting on you". Use full record
views only when detail is requested. For new intake, report the defined priority and queue
position. An explicit request for pending decisions follows the one-question protocol.
active_tasks groups recorded active lanes by task, so several workers on one task still
produce one task summary. running_without_lane is an ownership discrepancy to reconcile,
not proof that a process is dead. A lane record is not a live heartbeat. Label the channel of
a question waiting in another chat instead of presenting it as awaiting an answer here.
An idle status feed is opt-in through the project's notification policy, never enabled just
by installing this plugin. The executable Telegram status updater
provides configuration, preview, exchange recording, tick/loop and restart recovery. Integrate
it with the existing receiver's accepted-message and successful-send paths; do not run two
publishers for one bot/chat. If authorized and supported, use one quiet editable post with the
configured cadence/freshness limit; show a current user-facing sentence per active task, not
agent names. Keep public_summary and summary_updated_at on the canonical task, refreshing
them at dispatch and meaningful stage changes. Exclude inactive/stale summaries; do not add
a companion state file. Retire the post on a real exchange. Successful questions and valid
answers count as exchanges; typing indicators, the idle post itself and failed sends do not.
An empty, stale or failed read is not evidence that all work completed. The post never replaces
pending questions, task execution or receiver health checks.
Format for the actual channel: short heading, one fact per bullet, spacing between groups, and any decision last. Use rich text/HTML only when the installed bridge supports it; escape dynamic text with that channel's rules. Do not assume a transport flag or a project's language/format.
One presented question owns the decision channel until answered or explicitly parked; defer
unrelated topics and notification batches rather than making a bare numeric reply ambiguous.
Use coordinator-kit:question-protocol to correlate responses. A correction gets a brief
acknowledgment and an updated record, not a long apology or repeated interim conclusions.
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.
- 28d ago Changed · +20 lines 0c6e63594c5e
- 1mo ago Changed · +15 lines 0351cf0d10cb
- 1mo ago Changed · -12 lines · -172 tokens per session d67613eef550
- 2mo ago Changed · +3 lines bd69eb513b08
- 2mo ago First seen · 58 lines · 219 tokens per session scan A ba4049dc1a95
comms-register is a skill published in the GitHub repository nikolamin/claude-coordinator-kit (2 stars, last pushed 29d ago), licensed MIT. It adds 47 tokens to every session and 1,210 once invoked, about $0.0002 per session on Opus 5.5. 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.
comms-register compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| comms-registernikolamin/claude-coordinator-kit | Skill | 2 | 29d ago | MIT | A | 47 | Claude Code |
| imagegenopenai/codex | Skill | 134k | today | Apache-2.0 | A | 113 | Codex |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | today | MIT | A | 200 | Claude Code, Codex |
| playwright-climicrosoft/playwright | Skill | 99k | today | Apache-2.0 | A | 19 | Claude Code |
| systematic-debuggingobra/superpowers | Skill | 314k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 95 | Claude Code |
| chat-pet-sprite-creationmicrosoft/vscode | Skill | 195k | today | MIT | A | 53 | Claude Code, Codex |