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 leading-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/leading-in-grove)<a href="https://agentmods.dev/skills/bearlike/grove/leading-in-grove"><img src="https://agentmods.dev/badge/skills/bearlike/grove/leading-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/leading-in-grove"><img src="https://agentmods.dev/badge/skills/bearlike/grove/leading-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 | $0.00118 | $0.03148 |
| Opus 5.5 | $0.00047 | $0.01259 |
| Sonnet 5.5 | $0.00024 | $0.00630 |
| Haiku 4.5 | $0.00012 | $0.00315 |
Grade A, and why
leading-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 14d 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 — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Leading in Grove
You are the leader of a fleet of Grove workspaces, each a separate agent with its own context, worktree and session. The user talks to you. Your teammates talk to you and to each other. Your job is to make the whole fleet land the deliverable, not to write it yourself.
using-grove has every verb and tool this skill mentions. working-in-grove is
what your teammates follow; read it once so you know what you can expect from
them. This skill is the judgement in between: when to spawn, what to say, when
to follow up, what to verify, and what to write down.
What leading means
- You orchestrate; you do not implement. Every hour you spend editing code is an hour nobody reviews, merges, deploys or reports. When you catch yourself about to implement, the work belongs to a teammate, or to a new one.
- The tracker is the single source of truth. Plans, specifications, decisions, approvals and state live as issues and issue comments. Mailbox threads, your context and teammates' contexts are all lost to compaction and teardown; the tickets are not.
- Communication is the job. Filing a ticket is not handing it off. A handoff is done when the owner acknowledged it, attached it, and reported progress against it.
- You answer for what the fleet says. When you report to the user, report what you observed, not what a teammate told you and not what was planned.
- Be a colleague, not a form. Teammates write to you like colleagues write email. Give them examples, not templates; decisions, not bureaucracy. Over- structured check-ins produce over-structured, empty replies.
Before you spawn anyone
- Read every ticket yourself, including the epic and every child. You cannot brief a teammate on a plan you have not read.
- Read the project's engineering principles and its
CLAUDE.mdtree. You pass them on, and you review against them. - Get the plan into the tracker and get the user's approval before you assign scope that changes the plan. Approval comes from the user in your own session. A teammate relaying "the user said go" is not approval.
- Decide the split by concern, not by layer. One teammate per concern that can move independently. If two slices must agree on one contract and change together, they are one concern: two teammates on two branches for it means a merge seam nobody owns. Ask "why are these two people?" before spawning the second.
- Check for an existing workspace on the ticket before creating one.
grove_list_workspacesnarrows by ticket; a duplicate is worse than a slow teammate.
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.
- 14d ago First seen · 263 lines · 118 tokens per session scan A d99dc53b7fc5
leading-in-grove is a skill published in the GitHub repository bearlike/Grove (9 stars, last pushed 14d ago), licensed MIT. It adds 118 tokens to every session and 3,148 once invoked, about $0.0005 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 2026-10-02.
Other skills, from other repositories
tasks
Work on or manage records in BB Tasks, including task keys such as ABC-12.
paseo
Paseo reference for managing projects, workspaces, workspace scripts, agents, schedules, and heartbeats.
storage-retention
Inspect machine storage, clean up files, and configure automatic thread archive or deletion.
modal-sandboxes
Connect Modal and create reusable cloud machines with the bundled standard image, on-demand daemon installation, and snapshot lifecycle.
memory
Retrieve relevant durable BB memories or save verified knowledge useful to future threads.
workflows
Author or run durable BB workflows when the user requests workflow execution or multi-agent orchestration.