agentleFS
cacardinalcacardinal/laila/.claude/skills/new-domain/SKILL.md

new-domain

Skill1 starsChanged 4 months agoMITView on GitHub

Add or retire a life domain by executing EVERY step of the docs/adding-domains.md checklist — skeleton, registry, router, external systems, goals, then a verified trigger test. Use when Alex says "/new-domain", "add a domain", "new life domain", "spin up a domain for X", "retire the X domain", "kill the X domain", or "archive a domain". NOT /domain-hygiene (that scans existing domains for staleness; this one creates or retires them).

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 5 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Step A — Skeleton (Tier 1): Create:

    line 37
  2. Very specific:

    Step B — Registry (Tier 1): Add an entry to config/domain-triggers.json: patterns, context_file, state: "active", plus a meeting_routing …

    line 41
  3. Distinctive:

    Step C — Router pointer (Tier 3 — propose the diff): If it's a headline domain, prepare a one-line addition to the root AGENTS.md Domains …

    line 43
  4. Very specific:

    Step E — Goals: If the domain has quarterly goals, add them to the goal source of truth (the external system if one exists, else the …

    line 47
  5. Distinctive:

    External system: if one exists, propose marking the domain's record retired so dashboards stop counting it as live.

    line 61
  6. Very specific:

    Sweep for zombie references: search skills, scripts, and configs for the old domains/<name>/ path.

    line 63
  7. Very specific:

    Never summarize a partially-executed run as complete.

    line 80
  8. Distinctive:

    Before declaring this skill done (and again after edits):

    line 90
  9. Very specific:

    Add round-trip: add a throwaway domain (e.g.

    line 91
  10. Distinctive:

    If a step was skipped, mis-ordered, or a real add/retire later turned out partial (a zombie trigger, an empty goal tree), this skill is …

    line 97
  11. Very specific:

    If docs/adding-domains.md gains or changes a step, update this skill in the same session — the doc and the skill must never disagree.

    line 98

≈ “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

New Domain Skill

Operationalizes docs/adding-domains.md. Domains are cheap to add and cheap to retire — the failure mode is doing either partially: a router pointing at files that don't exist, or an archived domain whose trigger still fires. This skill exists so no step gets skipped. Read docs/adding-domains.md first every run; it is the checklist of record and this skill executes it.

Setup note: the registry is config/domain-triggers.json and the router is the root AGENTS.md — both must exist. Step 4 (external CRM/goal-tree registration) only applies if you run one; the skill asks rather than assumes. No scheduled jobs are created here — if a new domain later needs a background loop, that job uses the $AGENT_RUN env convention like every other loop.

Tier classification

  • Tier 1 (auto-execute + notify): creating files under domains/<name>/ and adding the registry entry to config/domain-triggers.json. Deterministic, reversible, invisible to others. Log to state/autonomy-audit.json, notify the command channel.
  • Tier 3 (propose-and-wait): any edit to the root AGENTS.md — it is the router every session loads, so present the exact diff and wait for approval. External CRM/goal-tree writes are also proposed first.
  • Retiring is always Tier 3. Moving a domain to the archive is destructive-adjacent; confirm the kill explicitly before touching anything.

Mode: Add (default)

/new-domain or /new-domain <domain-name>

1. Interview (brief — five questions, then move)

  1. Name — lowercase slug for the directory (domains/<name>/), display name for headers.
  2. What is it — one or two sentences: scope, what "checking" it should surface.
  3. Trigger phrases — at least "Check <name>"; ask for any natural variants Alex would actually say.
  4. Meeting-notes auto-filing — should meeting notes route here? If yes, collect the signals: participant email domains, people, company names, title keywords.
  5. Headline domain? — does it get a mention in the root AGENTS.md Domains line, or is it registry-only?

Also ask whether the domain has goals to seed and whether an external CRM/goal-tree system is in play (skip step 5 cleanly if not).

2. Execute — every step, in order

Track each step in a live checklist (see Reporting). A step is pending -> done -> verified; nothing ships at done.

Step A — Skeleton (Tier 1). Create:

  • domains/<name>/AGENTS.md — overview, a "Check <Name>" workflow (Read -> Detect -> Act -> Report, with Act split into Tier 1 vs Tier 3 actions), domain rules, trigger phrases. Model it on domains/ideas/AGENTS.md.
  • domains/<name>/tracking/status.md — with a **Last updated:** YYYY-MM-DD header (the domain-hygiene scanner keys off it), Current Focus, and an Active Items table. Model on domains/ideas/tracking/status.md.

Step B — Registry (Tier 1). Add an entry to config/domain-triggers.json: patterns, context_file, state: "active", plus a meeting_routing block (email_domains, people, company_names, title_keywords, filing.correspondence_dir) if the interview said notes should auto-file. Validate the JSON parses after the edit.

Step C — Router pointer (Tier 3 — propose the diff). If it's a headline domain, prepare a one-line addition to the root AGENTS.md Domains section (the active-domains list). Present the diff, wait for approval, then apply. The router routes — one line, no content.

Step D — External CRM/goal tree (Tier 3, only if Alex runs one). Propose the record creation AND the select-option additions on every object type that can be tagged with a domain (goals, tasks, KPIs). Surface the doc's warning verbatim in the proposal: the select-option step is the one that gets skipped — the domain appears in the system but nothing can be tagged to it, so its goal tree renders empty. When an API replaces an options array wholesale: read the full existing array, append the new option, write it ALL back — omitting an existing option deletes it.

Step E — Goals. If the domain has quarterly goals, add them to the goal source of truth (the external system if one exists, else the domain's status.md) and regenerate any exported views (state/goals.md).

Step F — Verify end-to-end. Have Alex (or a fresh session) say the trigger phrase and confirm: the registry resolves it, the domain's AGENTS.md loads, and the Check workflow runs against the new status.md. Only after this does the checklist read fully verified.

3. Close out

Log the Tier 1 actions to state/autonomy-audit.json, notify the command channel ("Domain <name> added — 6/6 steps verified"), and note the decision in knowledge/decisions/YYYY-MM.md.

Mode: Retire

/new-domain retire <domain-name> — all Tier 3. Confirm first: "Retiring <name>: archive the directory, kill the trigger, remove the router line. Confirm?"

  1. Archive the directory: git mv domains/<name>/ domains/_archive/<name>-killed-YYYY-MM-DD/ (plain mv if not tracked). Write ARCHIVED.md inside it: kill reason (Alex's words), last known state (pull from its status.md), and revival cost — what it would take to bring it back.
  2. Flip the registry: in config/domain-triggers.json, remove the trigger entry and append a row to the archived array (domain, state: "archived", kill date, archive path). Validate the JSON.
  3. External system: if one exists, propose marking the domain's record retired so dashboards stop counting it as live. Do NOT remove its select-option — historical records still reference it.
  4. Router: propose the root AGENTS.md diff removing the domain from the active list (and mentioning domains/_archive/ if it isn't already).
  5. Sweep for zombie references: search skills, scripts, and configs for the old domains/<name>/ path; fix or flag each hit.
  6. Verify: say the old trigger phrase in a fresh session — it must surface "that domain was killed <date>, archived at <path>" instead of running a Check workflow. A retirement that skips steps 2-4 leaves a zombie trigger that errors weeks later.

Reporting

Maintain and print the checklist at every pause point:

# Add domain: <name>
A. Skeleton (AGENTS.md + status.md)        [verified]
B. Registry entry (domain-triggers.json)   [verified]
C. Router pointer (root AGENTS.md)         [awaiting approval — diff above]
D. External CRM/goal tree                  [n/a — none configured]
E. Goals seeded + exports regenerated      [pending]
F. Fresh-session trigger test              [pending]

Never summarize a partially-executed run as complete. If the session must end mid-checklist, record the remaining steps in state/active-tasks.json (owner: laila) so the heartbeat nags until F is verified.

When to invoke proactively

  • Alex describes a new sustained commitment (a client engagement, a health program, a build) and starts asking Laila to track it — that's a domain forming; offer this skill instead of letting the tracking sprawl into daily notes.
  • A domain's status has read "nothing active" for a quarter and its Check runs come back empty — propose retirement rather than letting it rot (killed is a verdict, not a deletion; the archive keeps the post-mortem).
  • An engagement domain (the Acme pattern — created for one client commitment) hits the end of its contract: retirement is part of the close-out, not an afterthought.

Verification (MANDATORY)

Before declaring this skill done (and again after edits):

  • Add round-trip: add a throwaway domain (e.g. testbed), confirm all six steps reach verified including the fresh-session trigger test, then retire it via this same skill and confirm the trigger surfaces the archived state.
  • JSON integrity: config/domain-triggers.json parses after both the add and the retire edits.
  • Tier discipline: confirm the run never edited the root AGENTS.md without showing a diff and getting approval first.

Self-improvement (MANDATORY)

  • If a step was skipped, mis-ordered, or a real add/retire later turned out partial (a zombie trigger, an empty goal tree), this skill is wrong, not Alex. Propose the checklist edit immediately, or log {"skill": "new-domain", "correction": "...", "date": "YYYY-MM-DD"} to state/skill-feedback.jsonl.
  • If docs/adding-domains.md gains or changes a step, update this skill in the same session — the doc and the skill must never disagree.

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.