agentleFS
ByronScottJonesByronScottJones/Agents/skills/twg-responsibility-routing/SKILL.md

twg-responsibility-routing

Skill7 starsChanged 2 months agoMITView on GitHub

Use with root `twg` to route owners, subject-matter experts, maintainers, reviewers, approvers, decision authorities, Heads of Engineering, or escalation paths for a topic, area, ask, project, service, or component.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 2 are distinctive, and 6 are 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.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use the root twg skill. Use this workflow to answer “who should I involve, why them, and in what role?” Do not assume a broad topic is an …

    line 11
  2. Very specific:

    Run twg <command>. On shell command not found, use $HOME/.local/bin/twg (macOS/Linux) / $env:LOCALAPPDATA\Programs\twg\bin\twg.exe …

    line 16
  3. Distinctive:

    Keep these roles distinct:

    line 23
  4. Very specific:

    Use twg responsibility get <reference> for declared owners, teams, maintainers, approvers, reviewers, or escalation roles.

    line 48
  5. Very specific:

    Use twg responsibility infer <reference> only when declared responsibility is missing or the user asks for evidence-based candidates.

    line 54
  6. Very specific:

    When org data is connected and a candidate leader or manager is known, prefer the narrowest useful tree: org-tree for hierarchy, work-tree …

    line 60
  7. Very specific:

    Use twg-context-discovery only when relationship or dependency expansion is needed after the scope and people roles are established.

    line 73
  8. Distinctive:

    Stop when another search would not change scope, role, authority, escalation, confidence, or the recommended contact.

    line 90

≈ “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

twg-responsibility-routing

Use the root twg skill. Use this workflow to answer “who should I involve, why them, and in what role?” Do not assume a broad topic is an exact product.

CLI launcher fallback

Run twg <command>. On shell command not found, use $HOME/.local/bin/twg (macOS/Linux) / $env:LOCALAPPDATA\Programs\twg\bin\twg.exe (PowerShell), then tell user to add that directory to PATH. Do not treat auth or command errors as PATH failures.

Classify The Role

Keep these roles distinct:

  • Declared owner: formally accountable for an entity or area.
  • Operational owner or maintainer: responsible for current delivery or operation.
  • Expert: has source-backed depth in one workstream.
  • Reviewer or consulted party: evaluates or advises but may not decide.
  • Approver or decision authority: can authorize the specific ask or decision.
  • Escalation or Head of Engineering contact: leads the responsible engineering organization or can route an unresolved ask.
  • Informed stakeholder: affected by the decision but not responsible for it.

Resolve Scope Before People

For a fuzzy topic or area, compare one bounded twg rovo search "<topic>" --app confluence pass with one bounded twg work search "<topic>" pass. Group same-named results into distinct initiative, platform, domain, feature, and team clusters. Hydrate 2-5 central charters, roadmaps, services, projects, or active workstreams before ranking people. If the sources disagree, show the alternative scopes or ask the user.

For a stable project, goal, component, service, work item, page, or repository, use that reference directly.

Find Responsibility And Authority

  • Use twg responsibility get <reference> for declared owners, teams, maintainers, approvers, reviewers, or escalation roles. It takes an ARI, URL, or key, not a person's name.
  • For a person's manager, resolve the account with twg people search --name "<name>", then run twg user manager <account-id>. Never infer reporting lines from tickets or documents.
  • Use twg responsibility infer <reference> only when declared responsibility is missing or the user asks for evidence-based candidates. Preserve confidence, reason codes, evidence window, and declared-versus-observed status.
  • Infer primarily from scoped search, declared roles, and hydrated work evidence. Org data is optional; do not block or weaken the workflow merely because it is unavailable.
  • When org data is connected and a candidate leader or manager is known, prefer the narrowest useful tree: org-tree for hierarchy, work-tree for cross-surface breadth, workitem-tree for Jira work, or pr-tree for PR authorship/review. Tree counts help validate reach and organizational scope; they do not by themselves prove expertise, ownership, or approval authority.
  • For an approval ask, inspect the artifact that defines the decision: project or goal ownership, Jira workflow/approver fields, charter, decision record, or accountable team. State separately who can recommend, who must be consulted, and who can approve. Never promote an active contributor into an approver.
  • For escalation, give the nearest verified owner first. Add the engineering leader only when the owner is missing, blocked, cross-team, or the ask needs organizational authority.

Use twg-context-discovery only when relationship or dependency expansion is needed after the scope and people roles are established. It is not the default route for a responsibility question.

Evidence And Output

Strong evidence includes explicit ownership, accountable roles, workflow authority, charter/roadmap leadership, service ownership, sustained decisions, or repeated central contributions. Profile text, org proximity, one document, raw activity, or a self-authored update is weak evidence by itself.

For topic/SME maps, return at most five workstreams and one primary expert per workstream unless the user asks for more. Name an overall owner only with cross-workstream evidence. For approval or escalation asks, return the proposed contact, role, scope, evidence, confidence, why they are appropriate, and a suggested question or reach-out. State when authority is unconfirmed.

Stop when another search would not change scope, role, authority, escalation, confidence, or the recommended contact.

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.