ce-plan
The facts
Its description
Create structured plans for multi-step work, including software and non-software tasks. Use when asked to plan, break down implementation, plan from requirements, or deepen an existing plan; prefer ce-brainstorm for exploratory framing.
The description the agent reads to decide when to use it.
When it starts
The agent picks this when someone asks it to plan work: to plan a feature or task, break an implementation down, plan from a requirements document, or deepen an existing plan. The description sends exploratory framing to ce-brainstorm instead.
Things a person might say (the model's guesses)
- “Plan how to add pagination to the sessions API.”
- “Break this feature down into implementation steps.”
- “Deepen the plan at docs/plans/2026-05-12-pagination-plan.md.”
A model's summary of when it is used.
What it takes when started
[optional: feature description, requirements doc path, plan path to deepen, or any task to plan] [output:html]
What to type after its name when you start it.
Starts other agents
2 lines mention starting other agents
Lines in the file that mention starting another agent. Found by matching words, so treat them as pointers.
Files it mentions
references/output-contracts.md, references/output-mode.md, references/resume.md, references/intake.md (first 5 of 11 lines)
Files the text names, such as ones to read or write. Found by matching words, so treat them as pointers.
Stops for a person
2 lines stop to ask a person
Lines that stop and wait for a person before going on. Found by matching words, so treat them as pointers.
What it leads to
- hands off to ce-work plan-handoff.md, line 71
- calls ce-bakeoff only when a consequential HOW is left open bakeoff.md, line 3
- calls ce-doc-review only on the Durable path plan-handoff.md, line 7
- calls ce-noslop plan-sections.md, line 326
- calls ce-prototype only when a remaining feel-question is expensive to unravel plan-handoff.md, line 75
- an alternative: ce-brainstorm when the framing is exploratory line 3
- an alternative: ce-debug when the prompt is bug-shaped and the code is reachable intake.md, line 105
- an alternative: ce-optimize when the goal is optimization-shaped plan-sections.md, line 170
- an alternative: ce-pov when alternatives are already concrete enough to compare bakeoff.md, line 3
Other skills this one hands off to, calls, requires or offers instead, as a model read them.
What leads to it
- ce-brainstorm hands off to this when the user picks Create the implementation plan ce-brainstorm/handoff.md, line 107
- ce-doc-review hands off to this only when the user's existing request authorizes it ce-doc-review/synthesis-and-presentation.md, line 346
- ce-pov hands off to this only when the Phase 4 authority check passes ce-pov/boundaries.md, line 18
- ce-prototype hands off to this after a write-back, or when the session is enough to plan ce-prototype, line 59
- lfg calls this when the request changes code and no other route matches lfg, line 18
- ce-bakeoff requires this when ce-plan routes here ce-bakeoff, line 10
- ce-explain is an alternative to this only when implementation needs scoping ce-explain, line 80
- ce-work is an alternative to this for large cross-cutting work ce-work/work-intake.md, line 19
Other skills that hand off to this one, call it, require it or offer it instead, as a model read them.
Licence
MIT, from LICENSE
The terms its author shares it under, which let it be shown here.
Other files in this skill (35) Close
- references/approach-altitude.md
- references/bakeoff.md
- references/deepening-workflow.md
- references/final-review.md
- references/html-rendering.md
- references/intake.md
- references/markdown-rendering.md
- references/output-contracts.md
- references/output-mode.md
- references/plan-handoff.md
- references/plan-sections.md
- references/publishing.md
- references/reasoning-elevation.md
- references/research.md
- references/resume.md
- references/settled-decisions.md
- references/structure.md
- references/synthesis-summary.md
- references/universal-planning.md
- references/agents/agent-native-planning-strategist.md
- references/agents/architecture-strategist.md
- references/agents/best-practices-researcher.md
- references/agents/data-integrity-guardian.md
- references/agents/framework-docs-researcher.md
- references/agents/git-history-analyzer.md
- references/agents/learnings-researcher.md
- references/agents/pattern-recognition-specialist.md
- references/agents/performance-oracle.md
- references/agents/repo-research-analyst.md
- references/agents/security-sentinel.md
- references/agents/slack-researcher.md
- references/agents/spec-flow-analyzer.md
- references/agents/web-researcher.md
- scripts/elevation-dispatch.sh
- scripts/packs-resolve.py
The text
Most marks come from matching words in the file, so treat them as pointers.
Create Technical Plan
Outcome: a plan for carrying out and checking the agreed work while preserving its outcome and constraints. Resolve technical choices from evidence; leave adequate instructions unchanged. ce-brainstorm defines WHAT, ce-plan plans HOW, and ce-work executes. A prior brainstorm is optional.
An explicit invocation always produces a plan. Never classify a direct invocation as "not a planning task" and route out. It may select any output contract below, and the smallest valid plan is a few sentences in chat.
Research, decide, and write the plan — never implement. Do not write production code, run tests, or learn from execution-time results. Directional pseudo-code and grammar sketches may communicate design; changing code to see what happens belongs in ce-work.
Mandatory Completion Contract
A run is complete when its output contract's done condition is met. Every normal interactive branch that produces a plan artifact or checkpoint is incomplete until the user has been asked what to do next. A request that already authorizes the next action is that answer. For a Durable software implementation-plan run that continues past the resume check, that means the Phase 5.4 menu has been presented and the selected action has actually fired. For Direct, the change stated and the handoff offered; for a Chat brief, the brief and its one-line save-or-ce-work offer in chat. Neither presents the Phase 5.4 menu. Non-software plans and approach-level plans end with the handoff their reference workflow defines. A run that only answers a question may end after the answer unless its reference requires a save or share step.
Writing the file, checking confidence, and running or explicitly skipping ce-doc-review are intermediate milestones. In pipeline mode, the run is complete only when the plan, the confidence check, and the non-interactive document-review state have been returned to the caller, which decides what happens next.
Interaction Method
Ask one question at a time through the host's blocking question tool already in the current tool list. Match by capability; never probe a user-facing tool to discover it. If none is listed or a real question call errors, render numbered choices in chat; never silently skip a required question. If no feature description was supplied, ask what to plan and wait.
Raw lines 21 to 24Output Contract
Decide which output contract applies at the start of scoping (Phase 0.6), before choosing depth and before the scoping synthesis. It applies only when no resume route (Phase 0.1) fired and the source check (Phase 0.2) found no upstream artifact. Ground it with bounded inline reads of what the request names, without dispatching a subagent. Select one:
- Direct — the work can be stated, done, and verified in one pass with no decision the user would weigh. State the change in a few sentences and offer the handoff to
ce-workor the user; execution starts only with implementation authority, asreferences/output-contracts.mddefines. - Chat brief — bounded work with at most one decision the user would weigh and no risk surface. Deliver it in chat and stop.
- Durable — everything else. Continue the workflow below.
references/output-contracts.md defines Direct and Chat brief; read it when either is selected. When the tier is still uncertain after those reads, take the heavier one. If a read surfaces a decision the user would weigh, a risk surface, or multi-pass verification, move to the heavier tier before emitting anything. Durable regardless of size: a run with no synchronous user to act on chat this turn (pipeline, headless, goal- or scheduler-driven), a request whose wording asks for a plan, a plan file, or an output format, a request that continues an existing plan's item, and a risk surface — authentication, payments, migrations, external contracts.
Workflow
Phases run in order unless a reference routes out or short-circuits. Read a phase's required reference in full when you enter that phase; a read made before that phase does not satisfy it, and a reference named for the final steps is read again at its step even when already in context. If a required reference cannot be read, stop before the action it governs and report the blocker and recovery path; never reconstruct the missing rules from memory. In pipeline mode, every required-reference failure returns status: blocked, phase, blocker, and recovery_path; include artifact_path and preserve the artifact when one exists. Report blocked even when an artifact exists.
Phase 0: Output, Resume, and Scope
- Output first. Read
references/output-mode.mdbefore interpreting any phase. It defines token parsing, output and confirmation precedence, renderer selection, artifact location, and when a repository may be resolved. - Resume, deepen, approach, and domain. Read
references/resume.mdbefore acting. It defines resuming an existing plan, enriching a requirements-only plan, deepening, approach-level planning, and the software/non-software split. Follow any terminal route it selects; otherwise continue. - Source and scope. Read
references/intake.mdbefore Phase 0.2 and follow it through Phase 0.7. It defines finding and preserving the upstream artifact, routing out to bootstrap work, blocking questions, depth, named resources, and the scoping synthesis; the Output Contract decision above happens inside it. Do not pass a decision point that has not resolved.
Phases 1-4: Research and Compose
- Read
references/research.mdbefore gathering context. - Read
references/structure.mdbefore resolving questions or structuring the plan. - Compose from
references/plan-sections.mdplus the format-rendering reference selected byoutput-mode.md.
Phase 5: Review, Write, Deepen, and Hand Off
- Read
references/final-review.mdbefore the pre-write review; it defines Phase 5.1 through 5.3.2. On an interactive run from a brainstorm, the user sees the Phase 5.1.5 scoping synthesis in chat before the write: a confirmation to wait on, or its one-line auto-proceed announcement. - Model elevation. Immediately before authoring, read
references/reasoning-elevation.md, resolve the choice at this boundary, and follow it. Do not author until activation resolution has completed and any selected dispatch or transparent fallback has settled. - In pipeline mode, evidence that invalidates a decision settled earlier in the session stops the write. Return the exact token
settled-decision-invalidated, the decision, and the reason; do not resolve it silently. - Write the plan before presenting options, then complete the confidence path
final-review.mddefines.
**STOP. Read references/plan-handoff.md immediately before Phase 5.3.8 and 5.4** (document review and the handoff menu). Document review is mandatory for a Durable plan and the default is non-interactive (mode:non-interactive). In interactive software runs, ask exactly: "Plan ready at <absolute path to plan>. What would you like to do next?" Present its menu and wait. If the selection arrives after a user turn, reload references/plan-handoff.md before acting. Rendering the menu, receiving a selection, or announcing a route is not completion; execute the selected action.
From EveryInc/compound-engineering-plugin at 3449230, 2026-10-31. Skills read by DeepSeek V4.1 Flash on 2026-11-04. Summary and techniques by Claude Opus 5.5 on 2026-11-06.