agentleFS
redhat-developerredhat-developer/rhdh-skills/skills/reference/rhdh-spec-driven-schema/SKILL.md

rhdh-spec-driven-schema

Skill20 starsChanged 20 days agoApache-2.0View on GitHub

Owns RHDH's project-local OpenSpec workflow definition — the `rhdh-spec-driven` schema (proposal -> {specs, design} -> tasks -> apply), its four artifact templates, the Canonical Touchpoints rule that ties a change back to `specifications/prd/`, `specifications/adr/`, and `openspec/specs/<capability>/spec.md`, and the shared artifact-creation-loop mechanics every openspec-* skill drives through the `openspec` CLI. Also installs `config.yaml` and `schemas/rhdh-spec-driven/` into a product repo's `openspec/` so the CLI can resolve that schema. Invoked by name from openspec-new-change, openspec-continue-change, openspec-ff-change, and openspec-onboard; not a standalone entry point. Use for "what does the spec-driven schema require", "install the rhdh-spec-driven schema", "what goes in Canonical Touchpoints", "how do I fill in an artifact template", or "why did an artifact instruction reject my capability name".

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    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. Very specific:

    Run that from the product repo (or pass the project root as the first argument).

    line 52
  2. Very specific:

    Every proposal.md states a Canonical Touchpoints section naming every affected PRD/ADR file under specifications/ and every affected …

    line 74

≈ “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

RHDH spec-driven schema

Give every openspec-* skill one shared place to read the actual RHDH workflow definition, instead of each restating the schema's rules from memory.

What this skill owns

  • config.yaml — the project-local schema selection (rhdh-spec-driven), the RHDH context block (split canonical model, journal obligation), and the per-artifact house rules.
  • schemas/rhdh-spec-driven/schema.yaml — the authoritative artifact graph: proposal -> {specs, design} -> tasks -> apply, each artifact's instruction text, and the apply block's direct-vs-team mode guidance.
  • schemas/rhdh-spec-driven/templates/{proposal,spec,design,tasks}.md — the structural template for each artifact.
  • references/artifact-loop.md — the shared mechanics for driving openspec instructions <id> --change <name> --json and turning its response into a written artifact file.
  • scripts/install_project_schema.py — copies config.yaml and schemas/rhdh-spec-driven/ into a product repo's openspec/ so the OpenSpec CLI can resolve the schema there.

Install into the product repo (startup / setup)

OpenSpec loads schemas from the project's openspec/, not from this skill directory. Before any openspec new change that should use rhdh-spec-driven, ensure those files are on disk:

python3 scripts/install_project_schema.py

Run that from the product repo (or pass the project root as the first argument). The helper lives next to this skill — resolve scripts/install_project_schema.py relative to this skill's install path, not under the product repo's scripts/. Use --force only when deliberately replacing a customized copy.

Callers (/openspec-new-change, /openspec-ff-change, /openspec-onboard, and /setup-rhdh-skills when seeding a product checkout) check for openspec/config.yaml and openspec/schemas/rhdh-spec-driven/ first and run this install step only when either is missing. Idempotent: existing files are kept unless --force is set. Writing into the product repo is an external write — take it through /mutation-gate when the caller is in a setup or multi-operation plan; a single scaffold turn that already creates openspec/changes/ may include this copy in the same stated set.

Confirm with openspec schemas --json that rhdh-spec-driven is listed, then omit --schema to take the configured default (or pass --schema rhdh-spec-driven explicitly).

Canonical Touchpoints, non-negotiable

Every proposal.md states a Canonical Touchpoints section naming every affected PRD/ADR file under specifications/ and every affected long-lived capability spec under openspec/specs/, or explicitly None, plus the change type: product | architecture | feature-spec | migration | workflow-only | docs-only. design.md and tasks.md carry that same touchpoint set forward — see schema.yaml's per-artifact instruction field for the exact wording each artifact requires. A caller skipping this because "it's a small change" is exactly the case the rule exists for: state None explicitly rather than omitting the section.

Journal obligation

config.yaml's context block states the turn-bookending discipline (log turn.start before work, turn.end after, every prompt, inside an active change) and the apply-phase event set. The mechanics of writing those events belong to openspec-journal, invoked by name — this skill states when the obligation applies; openspec-journal states how to satisfy it.

Completion

Complete when the caller has either installed openspec/config.yaml and openspec/schemas/rhdh-spec-driven/ into the product repo, or read the exact schema/template/context text it needed for the artifact in front of it — rather than guessing at wording. A caller citing this skill without reading schema.yaml's instruction field for that artifact is the failure mode this skill exists to prevent.

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.