agentleFS
imMamdouhaboammarimMamdouhaboammar/get-things-done/skills/building-gtd-domain-packs/SKILL.md

building-gtd-domain-packs

Skill2 starsChanged 3 months agoMITView on GitHub

Use when someone says "GTD doesn't cover my field well", "extend GTD for legal/medical/design/data work", "build a domain pack for X discipline", "create a GTD extension without forking the core", or "customize GTD for my industry". Creates a focused domain extension that inherits Get Things Done instead of copying it. Do not use for creating a new agent skill unrelated to GTD, using GTD to execute work (use get-things-done instead), or adding a feature to GTD's core contract.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 9 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create a focused domain extension that inherits Get Things Done instead of copying it.

    line 9
  2. Distinctive:

    Write positive and negative selection signals based on intent and artifacts, not isolated keywords

    line 22
  3. Distinctive:

    Keep terms whose confusion would change decisions, execution, or verification

    line 25
  4. Very specific:

    Put optional field data under domain_data.

    line 31
  5. Very specific:

    Use templates/domain-pack-template.md as the output shape.

    line 45
  6. Distinctive:

    Do not copy the core workflow into the domain pack

    line 49
  7. Distinctive:

    Do not turn field knowledge into a questionnaire

    line 50
  8. Distinctive:

    Do not create a new domain for branding or naming alone

    line 51
  9. Distinctive:

    Do not weaken decision authority, readiness, evidence, completion, handoff, or tool honesty

    line 52
  10. Distinctive:

    Do not claim a domain is validated until the four collision cases were actually exercised

    line 53
  11. Distinctive:

    Before calling the domain pack ready:

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Building GTD Domain Packs

Create a focused domain extension that inherits Get Things Done instead of copying it.

Load references/domain-pack-spec.md at the start — the required headings, field contracts, and validation rules live there. Load references/core-contract.md to confirm what the core already handles before extending it.

Build method

  1. Collect failure examples first Gather at least four realistic tasks: messy in-domain, well-formed in-domain, deceptive near-complete, and adjacent out-of-domain
  2. Prove the core is insufficient Identify which field-specific vocabulary, readiness, review, or completion behavior is missing. If the core already handles the examples, do not create a pack
  3. Define selection boundaries Write positive and negative selection signals based on intent and artifacts, not isolated keywords
  4. Extract only consequential vocabulary Keep terms whose confusion would change decisions, execution, or verification
  5. Add diagnostic candidates Include only questions that can change material work. Do not make them mandatory
  6. Extend the brief safely Put optional field data under domain_data. Never fork core fields
  7. Add stricter gates where needed Define domain-specific readiness checks, 2 to 4 workstream patterns, specialist review checks, and observable completion evidence
  8. Test routing collisions The adjacent out-of-domain case must remain outside this pack. Tighten selection signals until it does
  9. Test deceptive completion A plausible deliverable without field-specific proof must not pass Done
  10. Validate the contract Confirm every required heading appears exactly once and the pack does not weaken the core

Use templates/domain-pack-template.md as the output shape.

Rules

  • Do not copy the core workflow into the domain pack
  • Do not turn field knowledge into a questionnaire
  • Do not create a new domain for branding or naming alone
  • Do not weaken decision authority, readiness, evidence, completion, handoff, or tool honesty
  • Do not claim a domain is validated until the four collision cases were actually exercised

Self-validation checklist

Before calling the domain pack ready:

  • all 9 required headings present exactly once
  • YAML frontmatter has domain, version, and extends: gtd-core-v1
  • selection signals include at least one non-selection signal
  • vocabulary section explains why confusion changes work (not just a glossary)
  • diagnostic questions are candidates, not a mandatory questionnaire
  • readiness additions are observable by another agent
  • 2–4 workstream patterns with dependency edges
  • completion checks prefer inspectable artifacts over subjective language
  • four collision cases tested (in-domain messy, in-domain well-formed, deceptive near-complete, adjacent out-of-domain)
  • adjacent out-of-domain case does NOT route to this pack

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.