Source profileQuality 91/100

OpenDigitalProductFactory/opendigitalproductfactory/packages/dpf-skill-pack/skills/dpf-writing-plans/SKILL.md

dpf-writing-plans

Use when a filed DPF backlog item needs an implementation plan before code is written.

Source repository stars
14
Declared platforms
0
Static risk flags
0
Last source update
2026-08-28
Source checked
2026-08-28

Decision brief

What it does: where it fits

When work is bigger than a single obvious edit, write the plan before the code — but in DPF a plan is not a floating document. It hangs off a filed backlog item. This skill is the DPF-native planning step; it replaces the retired upstream writing-plans skill so the flow resolves…

Best for

  • A filed BI describes a multi-step build, migration, or refactor with ordering constraints.
  • A spec is approved and you need the implementation sequence before touching code.
  • The operator says "plan out X" and X is more than a one-file change.

Not for

  • The work is a single obvious edit — just do it; a plan is overhead.
  • There is no BI yet — file one first (dpf-file-backlog-item); a plan for floating intent is the substrate violation this skill exists to prevent.

Compatibility matrix

Platform support, with evidence labels

PlatformStatusEvidenceWhat to check
CodexNot declaredNo explicit evidencePortability before use
Claude CodeNot declaredNo explicit evidencePortability before use
CursorNot declaredNo explicit evidencePortability before use
Gemini CLINot declaredNo explicit evidencePortability before use
Open the compatibility checker

Installation

Inspect first. Install second.

The source command is displayed only when detected. A safe inspection prompt is always available so your agent can explain every action before execution.

Source-detected install commandSource
npx skills add https://github.com/OpenDigitalProductFactory/opendigitalproductfactory --skill "packages/dpf-skill-pack/skills/dpf-writing-plans"
Safe inspection promptEditorial

Inspect the Agent Skill "dpf-writing-plans" from https://github.com/OpenDigitalProductFactory/opendigitalproductfactory/blob/f795645f8fda2da7c172b32be5014f37517cee10/packages/dpf-skill-pack/skills/dpf-writing-plans/SKILL.md at commit f795645f8fda2da7c172b32be5014f37517cee10. List every install step, command, network request, credential, file read/write, external action, and rollback step. Explain whether it fits my task. Do not install or execute anything until I approve.

Workflow

What the source asks the agent to do

  1. 01

    When to use

    A filed BI describes a multi-step build, migration, or refactor with ordering constraints.

    A filed BI describes a multi-step build, migration, or refactor with ordering constraints.A spec is approved and you need the implementation sequence before touching code.The operator says "plan out X" and X is more than a one-file change.
  2. 02

    When NOT to use

    The work is a single obvious edit — just do it; a plan is overhead.

    The work is a single obvious edit — just do it; a plan is overhead.There is no BI yet — file one first (dpf-file-backlog-item); a plan for floating intent is the substrate violation this skill exists to prevent.The approach itself is unsettled — brainstorm/decide first (dpf-brainstorming → dpf-decision-via-kernel), then plan the chosen approach.
  3. 03

    Steps

    1. Anchor to the BI. Confirm the BI exists and read it (mcpdpfgetbacklogitem). The plan implements that BI; its acceptance criteria are the plan's definition of done.

    Anchor to the BI. Confirm the BI exists and read it (mcpdpfgetbacklogitem). The plan implements that BI; its acceptance criteria are the plan's definition of done.Ground in the substrate before sequencing. Don't plan against an imagined codebase:mcpdpfsearchspecsandplans — an approved design or prior plan may already cover this.
  4. 04

    Guardrails

    No plan without a BI. A plan is for a filed BI, not a TODO or a floating doc.

    No plan without a BI. A plan is for a filed BI, not a TODO or a floating doc.No independent work only in Markdown. Planning is incomplete until each independently shippable deliverable has a live BI mapping and the plan carries a valid receipt.No silent MCP bypass. Unreachable MCP, a missing tool, or insufficient token scope stops planning before source implementation.

Permission review

Static risk signals and limitations

No configured static risk pattern was detected

This is not proof of safety. Runtime behavior, indirect dependencies, and hidden external systems are outside the static scan.

Evidence record

Why each signal appears

EvidenceSourceComputedTestedEditorial
SignalValueEvidence typeMeaning
Quality score91/100ComputedDocumentation, specificity, maintenance, and trust rules
Repository stars14SourceRepository attention, not individual Skill quality
Compatibility0 platformsSourceDeclared in the catalog source record
Usage guideautomated source guideEditorialGenerated or reviewed according to the visible evidence level

Pinned source

Provenance and original SKILL.md

Repository
OpenDigitalProductFactory/opendigitalproductfactory
Skill path
packages/dpf-skill-pack/skills/dpf-writing-plans/SKILL.md
Commit
f795645f8fda2da7c172b32be5014f37517cee10
License
Apache-2.0
Collected
2026-08-28
Default branch
main
View the original SKILL.md

DPF Writing Plans

When work is bigger than a single obvious edit, write the plan before the code — but in DPF a plan is not a floating document. It hangs off a filed backlog item. This skill is the DPF-native planning step; it replaces the retired upstream writing-plans skill so the flow resolves from one DPF source on every surface.

The order is fixed: BI first, then plan. dpf-file-backlog-item creates the governed work record; this skill turns it into an executable, phased plan.

When to use

  • A filed BI describes a multi-step build, migration, or refactor with ordering constraints.
  • A spec is approved and you need the implementation sequence before touching code.
  • The operator says "plan out X" and X is more than a one-file change.

When NOT to use

  • The work is a single obvious edit — just do it; a plan is overhead.
  • There is no BI yet — file one first (dpf-file-backlog-item); a plan for floating intent is the substrate violation this skill exists to prevent.
  • The approach itself is unsettled — brainstorm/decide first (dpf-brainstormingdpf-decision-via-kernel), then plan the chosen approach.

Steps

  1. Anchor to the BI. Confirm the BI exists and read it (mcp__dpf__get_backlog_item). The plan implements that BI; its acceptance criteria are the plan's definition of done.

  2. Ground in the substrate before sequencing. Don't plan against an imagined codebase:

    • mcp__dpf__search_specs_and_plans — an approved design or prior plan may already cover this.
    • mcp__dpf__search_code_graph — find the real files/contracts the plan will touch.
    • This is the consult-specs-first + research-before-implementing discipline applied to planning.
  3. Consume the start-of-work impact contract. Once exact edit paths are claimed, read the Workroom's verificationState.changeImpactContract (also returned by claim_capsule_scope). Put every testImpact and guardObligation in the relevant phase before implementation begins. If the contract is unresolved, plan exhaustive verification and an explicit impact-resolution step; never translate missing advice into an exemption.

  4. Write phased steps and classify delivery boundaries. Each phase needs a concrete deliverable, touched files, dependencies, and functional verification. Mark whether it is independently shippable or only internal sequencing. For every independently shippable deliverable, query for an existing covering BI and reuse it when present; otherwise compose with dpf-file-backlog-item and file it. An xlarge umbrella always requires this decomposition decision before implementation.

    If the plan opens with an agent-execution preamble, use the DPF-native one — never copy the retired superpowers:* boilerplate from older plans (~280 historical plans still carry it; those skills no longer exist on any surface, and the check-no-retired-superpowers-skills CI ratchet fails a new reference). Canonical preamble:

    For agentic workers: execute this plan one independently reviewable backlog item at a time — one BI, one branch, one PR. Use dpf-tdd for red-green implementation, dpf-local-merge-ci-before-push plus the plan's completion gate before any success claim, and dpf-pr-with-dco for handoff.

  5. Record live backlog coverage. Call mcp__dpf__record_plan_backlog_coverage with the umbrella BI, plan path, deliverable graph, BI mappings, and decision:

    • decomposed: every independent deliverable maps to a live new or existing BI;
    • atomic: no deliverable is independently shippable, with a substantive operator rationale explaining why one BI is correct.

    Copy the returned receipt, parent BI, deliverable-to-BI mappings, and dependencies into a ## Backlog coverage plan section. mcp__dpf__check_plan_backlog_coverage is the resumability check. If MCP is unavailable, the tool is missing, or the token lacks scope, stop and report that condition; Markdown checkboxes are not a fallback and planning/backlog completeness cannot be claimed.

    Treat a rejected coverage write as a remediation contract, not a blind retry:

    • When the result is traceability-incomplete, add or repair the initiative baseline and mappings named by the response before making one corrected call. Do not repeat identical arguments.
    • When the result is plan-artifact-invalid, reconcile the Workroom to the exact current branch and head SHA, repair the artifact or provenance named by the response, then make one corrected call. Do not mint a new plan path to dodge immutable provenance.
    • When the response says retryable: false, stop after following any explicit remediation that changes the request. If no authorized remediation is available, report the returned blocker; repeated calls cannot make the contract valid.
  6. Name the risks and the rollback. What could break (blast radius), and how to back out. A plan that only describes the happy path is half a plan.

  7. Save it. docs/superpowers/plans/YYYY-MM-DD-<feature-name>.md, cross-referencing the BI id and the coverage receipt. This path is where search_specs_and_plans and reviewers expect plans to live. Format is opt-in: Markdown is the default and stays fully supported, but when the plan leans on a flow/state diagram, a multi-column table, or side-by-side option fan-out, an HTML artifact often reads better and keeps the operator in the loop — see html-artifacts-guide.md and the _templates/spec.template.html starting point. If you ship HTML-only, leave a short Markdown stub carrying the canonical coverage section so search_specs_and_plans and the guard can find it.

Guardrails

  • No plan without a BI. A plan is for a filed BI, not a TODO or a floating doc.
  • No independent work only in Markdown. Planning is incomplete until each independently shippable deliverable has a live BI mapping and the plan carries a valid receipt.
  • No silent MCP bypass. Unreachable MCP, a missing tool, or insufficient token scope stops planning before source implementation.
  • No plan against an unverified codebase. Ground every "touch file X" claim in a real grep / code-graph hit.
  • Every phase carries its verification. A phase with no "how I'll know it works" is an aspiration, not a step.

See also

Frequently asked questions

What to verify before installation and use

What does the dpf-writing-plans source document cover?

When work is bigger than a single obvious edit, write the plan before the code — but in DPF a plan is not a floating document. It hangs off a filed backlog item. This skill is the DPF-native planning step; it replaces the retired upstream writing-plans skill so the flow resolves…

How do I install dpf-writing-plans?

The source record exposes this install command: npx skills add https://github.com/OpenDigitalProductFactory/opendigitalproductfactory --skill "packages/dpf-skill-pack/skills/dpf-writing-plans". Inspect the command and pinned source before running it.

Alternatives

Compare before choosing

Computed 10045,960

coreyhaines31/marketingskills

ab-testing

When the user wants to plan, design, or implement an A/B test or experiment, or build a growth experimentation program. Also use when the user mentions "A/B test," "split test," "experiment," "test this change," "variant copy," "multivariate test," "hypothesis," "should I test this," "which version is better," "test two versions," "statistical significance," "how long should I run this test," "growth experiments," "experiment velocity," "experiment backlog," "ICE score," "experimentation program

Computed 1008

narrative-io/narrative-skills-marketplace

design-analysis

Translate a fuzzy analytical question into a rigorous investigation plan. Interrogates the ask, grounds the plan in the available data dictionary, applies analytical best practices, and produces a structured brief of query specifications for a downstream query-writing skill. Plans, does not write SQL. Use when: "why did X drop", "is there a relationship between A and B", "who are our highest-value customers", "what's driving the change in Y", "investigate this trend", "design an analysis for", "

Computed 9967

brucesongs/kali-claw

insecure-design

Insecure Design (OWASP A06:2025) focuses on security flaws in system architecture and design phases, rather than code implementation-level bugs.

Computed 9916

NintendaDev/unikit-ai

unikit-docs

Generate and maintain the project's TECHNICAL documentation from its codebase — scans the project structure, tech stack, and module boundaries, then writes a lean README landing page plus detailed topic pages (architecture, modules, setup, build, APIs), only the docs that are relevant. Use whenever the user wants to create, update, or validate documentation of the CODE or the project itself, e.g. "generate documentation", "create docs", "write the README", "update the project docs", "document th