Skip to content
Agentmart
Submit a tool

bearing-lite

Coordinate a Bearing lifecycle as Planning or Implementation Orchestrator only when the owner explicitly requests Bearing for that mode.

Skill by alphazede

IndexedApache-2.0★ 3
Write a review

Chalk Labs

Need help deploying bearing-lite?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/alphazede-bearing-lite

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Coordinate a Bearing lifecycle as Planning or Implementation Orchestrator only when the owner explicitly requests Bearing for that mode. Do not activate for ordinary work, model or role selections, assigned packets, implementation work without that request, or advisory-hook suggestions.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 26, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/bearing-lite

What's inside

Skill name
bearing-lite
Description
Coordinate a Bearing lifecycle as Planning or Implementation Orchestrator only when the owner explicitly requests Bearing for that mode. Do not activate for ordinary work, model or role selections, assigned packets, implementation work without that request, or advisory-hook suggestions.
License
Apache-2.0
Instructions
53 lines in SKILL.md · view on GitHub

Files (16)

  • SKILL.md
  • openai.yaml · agents/
  • assurance-policy.md · references/
  • lineups.md · references/
  • owner-stops.md · references/
  • peer-synthesis.md · references/
  • planning-assembly.md · references/
  • profiles.md · references/
  • resume.md · references/
  • review-capability.md · references/
  • review-policy.md · references/
  • role-routing.mmd · references/
  • task-state.md · references/
  • task-state.mmd · references/
  • verification.md · references/
  • task.md · templates/

SKILL.md (first lines)

# Bearing Orchestrator

Coordinate planning, implementation, verification, validation and delivery. Preserve scope, agents, authority and evidence.
Own lifecycle state and owner communication; unrequested review stays inactive. Resolve Coordinator escalations using `references/owner-stops.md#coordinator-escalation-and-delegation`; delegation defaults off and requires scoped authority.
After opt-in, bind the session/task in the project plan per `templates/task.md`; installation, old plans, or role names never activate hooks. On resume, follow `references/resume.md` first; continue the recorded mode and next unfinished action.
## Modes

- **Planning Orchestrator:** owns context, specialist dispatch, final plan/design, owner agreement, and the written handoff; does not dispatch implementation.
- **Implementation Orchestrator:** consumes the handoff, checks authority/candidate, dispatches enabled roles, and owns delivery.

1. **Intake:** use `../intake/SKILL.md` to confirm repository and task location. Reuse confirmed inputs; ask only on material ambiguity.
2. **Architectural Alignment and Scope:** use `../architectural-alignment/SKILL.md`
   in the controller session to establish architecture, constraints, and shared
   scope. Own the workspace and map; add no second scope stage or agent.
3. **Planning:** coordinate profile-enabled specialists, then author final plan/design
   and assemble the package using `references/planning-assembly.md`. Run requested
   planning review after assembly, before owner confirmation/handoff.
4. **Implementation:** consume the handoff; in selected TDD run the frozen Test Implementer (`roles.test_implementer`)
   before Product Implementer. Send `work_class: light` slices to a selected Light Implementer; `reclassify: judgement`
   returns them to Implementer. Honor dependencies, commands, and write sets; use
   `../coordinator/SKILL.md` only for a justified profile-enabled coordination role.
5. **Review:** dispatch owner-requested review immediately when each frozen task, phase,
   or lifecycle boundary passes deterministic checks; never re-ask or defer it. An independent evaluator is a reviewer.
   Preserve selected assistance tools; preauthorize bounded same-session repair in the
   original packet and consume consolidated evidence without per-finding approval. Add no quality agents.
6. **Landing and closeout:** land each closed unit (`references/task-state.md#landing-a-closed-unit`); check deliverables and evidence. Distinguish
   author checks, requested review, merge, deployment, and activation; report remaining obligations.

## Planning handoff

Write `handoff.md` using the Prompt skill (`../prompt/SKILL.md`) in the confirmed plan directory when planning is complete. Include exact plan paths and digests, scope, owner decisions and authority,
selected models/harnesses/capabilities, candidate and worktree, assignments and write sets, commands, evidence, review status, blockers, and the next action.
Recommend a new Implementation Orchestrator session with this handoff; never launch it automatically. Continuing here is an owner choice.
The implementation session consumes settled planning; it does not repeat it.
Handoff and closeout pass `journey` to `hooks/closeout.cjs`; untranscribed decisions, authority events, or landings refuse them until the configured recorder transcribes them.
## Dispatch and recovery

Preserve the recorded profile; use `references/profiles.md` for authorized amendments.
Launch each selected agent assignment once. Extra invocations, retries, repairs,
or resumed agent runs require explicit authority; self-correction stays in-run.
An enabled Scribe uses one session at its frozen cadence (`event` default); never relaunch it per event or phase. When disabled or omitted, Orchestrator records history.
Register native completion notifications at dispatch. Wait through the harness's event mechanism; wake when an agent response arrives.
Consume each response once and dispatch newly ready work. Never poll logs, processes,
status, or worktrees, or use timers/sleep loops for periodic “still running” checks.
If event waiting is unavailable, return `WAITING_ON`; resume on the actual response.
Use `references/task-state.md` for state integrity, `references/resume.md` for resume and Scribe coverage, and `references/owner-stops.md` for authority/intent stops. Never replay completed stages or change routes.
Selected or required capabilities activate; unavailability, including a needed Coordinator or Systems Modeler omitted or disabled, is a typed capability gap, not substitution, not success, not invented behavior. Unselected and unrequired absence remains inactive, not a global failure.
Return `READY`, `WAITING_ON`, `OWNER_DECISION_REQUIRED`, or `COMPLETE` with mode,
stage, candidate, completed work, evidence, blocker, and next action.
Never expand scope, author delegated artifacts, invent decisions, or pass missing evidence.

From alphazede/bearing-lite, Apache-2.0.

More from alphazede/bearing-lite

Signals

Stars
3
Forks
1
Open issues
9
Last commit (this folder)
Oct 18, 2026 (12 days ago)
Last release
Oct 21, 2026 (9 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Apache-2.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from alphazede

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get bearing-lite running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying bearing-lite?

CallGet help
Cite this page

bearing-lite skill. Agent Mart. https://agentmart.dev/skills/alphazede-bearing-lite (last updated 2026-10-24)