agentleFS
alphazedealphazede/bearing-lite/skills/planning-and-design/SKILL.md

planning-and-design

Skill3 starsChanged 3 months agoApache-2.0View on GitHub

Perform final planning/design in the Planning Orchestrator's current session after the selected specialists return. Use for final plan/design and assembly or an authorized planning correction. Do not dispatch a separate author, repeat specialist requirements or SEIT work, or start implementation.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Verify settled intent and source/candidate bindings.

    line 25
  2. Distinctive:

    Do not draft a duplicate requirements register or SEIT.

    line 33
  3. Very specific:

    Check your own final package: run the register precheck at its final stage with --seit, every pending binding completed, then …

    line 43

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Final Planning and Design

Controller procedure, not a configured specialist role or extra invocation. Use ../bearing-lite/SKILL.md and its planning-assembly instructions.

Inputs and loop

Consume confirmed alignment/scope, requirements and definitions, architecture/model inputs, boundary and assembly contracts, SEIT, evidence, selected routes/capabilities, cadence, authority, and repository rules. Reuse adequate sources for inactive roles; expose missing required inputs rather than inventing them. Any unresolved material scope, behavior, authority, risk, or acceptance intent returns REROUTE_SCOPE_DEFINITION to combined alignment/scope in this controller session; generate no implementation.json or Manifest and never choose for the owner.

  1. Verify settled intent and source/candidate bindings. Establish whether a requirements register exists; if evidence cannot decide, ask the owner. Never infer one; keep register references versus Lifecycle-local requirements distinct.
  2. Author the technical plan and design.md from the completed inputs. Preserve registered IDs and definitions, published-standard clauses, trust boundaries, and recovery. Copy the confirmed quantity-definitions table into the glossary verbatim (references/artifact-grammar.md). Allocate requirements to components from the placement receipt per references/allocation-inputs.md. Author the OOPDSA Implementation Design section per references/oopdsa-design.md; plan programs per references/program-planning.md. Do not draft a duplicate requirements register or SEIT.
  3. Build coherent bounded slices and dependency-based waves with exact write sets, acceptance, commands, shared resources, and serialization reasons. Tests meet ../test-engineer/references/shared-machine-tests.md; engine-backed cases carry resource_group/expected_gate_seconds, waves added_gate_seconds/resource_groups. Prioritize prerequisite work and parallel ready work; avoid overloaded batches.
  4. Assemble implementation.json and the DoD Manifest. Copy specialist-owned decisions faithfully per ../bearing-lite/references/planning-assembly.md; choose no new roles or capability defaults. Record planning_wave with the declared enabled sessions and review_requested.
  5. Check your own final package: run the register precheck at its final stage with --seit, every pending binding completed, then plan-package.cjs draft. Repair your owned artifacts in-session and rerun affected checks. Missing specialist facts stay gaps; return them to their owner, never author them yourself.
  6. Requested planning review opens only after hooks/planning-review.cjs evaluatePlanningReview() passes its guards (manifest_not_generated without the rendered Manifest). Then obtain one integrated owner confirmation (never insert a profile or route pause), run the final freeze, write the handoff, and recommend a new Implementation Orchestrator session.

Return and recovery

Return READY, REROUTE_SCOPE_DEFINITION, WAITING_ON, or OWNER_DECISION_REQUIRED with candidate, paths, checks, open items, and next action. An authorized named delta packet applies named findings only, through references/correction-loop.md and hooks/correction-delta.cjs, within its existing authority and allowance; specialist-owned rows return to their author. This controller author check is not independent review or an extra agent pass.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.