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 bearlike/Grove --skill working-in-grovegit clone --depth 1 https://github.com/bearlike/GroveWrote 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/bearlike/grove/working-in-grove)<a href="https://agentmods.dev/skills/bearlike/grove/working-in-grove"><img src="https://agentmods.dev/badge/skills/bearlike/grove/working-in-grove/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/bearlike/grove/working-in-grove"><img src="https://agentmods.dev/badge/skills/bearlike/grove/working-in-grove.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 | $0.00100 | $0.03676 |
| Opus 5.5 | $0.00040 | $0.01470 |
| Sonnet 5.5 | $0.00020 | $0.00735 |
| Haiku 4.5 | $0.00010 | $0.00368 |
Grade A, and why
working-in-grove 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 23d 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 — 328 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Working inside a Grove workspace
You are one agent in a fleet. Somebody is watching twenty workspaces at once and cannot attach to yours to find out how it is going. Report your task phase so they can tell scope from handoff at a glance.
Writing to another agent
Every live Grove agent on this host can be written to, and you are one of them. This is true whether an agent runs as a Grove-owned native session (the default for Claude Code, Codex and OpenCode) or in its own interactive terminal. There is nothing to enrol in and no credential to obtain.
Find out who is reachable:
grove mailbox contacts
Each contact carries the address you write to, who it is, and whether it is
live right now. A contact that is not live is paused, offline or still
building; write to it after it comes back rather than retrying.
Send one message:
grove mailbox send --to <workspace-id> --subject "Ready for review" --body "PR #42 is open."
--from defaults to your own workspace, read from GROVE_PHASE_FILE. Add
--to-agent <slot> for a named agent sharing a container, --from-agent to say
which slot you are, and --body-file - to pipe a long body through stdin.
A reply is the same command with the two addresses swapped. Every message
you receive states the sender's address and the exact command that answers it,
so a conversation never depends on a receipt still being valid. Pass
--in-reply-to <message-id> to let the reader follow the thread.
The MCP equivalents are grove_list_mailbox_contacts and
grove_send_mailbox_message, registered on every Grove MCP server.
What a receipt does and does not say. delivered means Grove handed the
text to that agent's session — not that a model read it, agreed with it, or
acted on it. rejected names a reason you can act on (not_live, too_large);
unknown means the transport neither confirmed nor refused, so the message may
or may not have landed. Do not retry a send, type into somebody's pane, or treat
a successful tool call as compliance.
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.
- 23d ago Changed · +11 lines 08c77c2e6112
- 1mo ago Changed · +49 lines · +14 tokens per session 408d5f79e24c
- 1mo ago Changed · +1 lines · -9 tokens per session bfdd651e5b9d
- 1mo ago Changed · +51 lines · +95 tokens per session 1479c484ba09
- 1mo ago First seen · 216 lines · 0 tokens per session scan A ed2f0ee555a3
working-in-grove is a skill published in the GitHub repository bearlike/Grove (9 stars, last pushed 24d ago), licensed MIT. It adds 100 tokens to every session and 3,676 once invoked, about $0.0004 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.
Other skills, from other repositories
pipeline-conductor
Operating procedure for the kirocrew-pipeline-conductor agent. Use when a pipeline conductor session is seeded, inspected or debugged. You MANAGE one issue/PR pipeline on one repo: two columns per item (session working? item solved?), send intent, never re-derive a red yourself.
tasks
Work on or manage records in BB Tasks, including task keys such as ABC-12.
github-pr-status
Inspect an environment's GitHub pull request, mark it ready or draft, or merge it from the bb CLI; read PR status from another plugin.
github
Drive GitHub via the official gh CLI — repos, issues, pull requests, releases, gists, Actions runs, and raw REST through gh api. Use when the user asks to inspect or manage GitHub.
orchestrate
Toolbox for driving HumanLayer from the CLI - log in, pick a host, launch tasks in their own worktrees, launch sessions, list tasks and their PRs, and copy, move, or upload task artifacts.
repo-conventions
Repository conventions for commits, pull requests, and tests in this example project.