Openspec Aware

OpenSpec-aware Authoring (OpenClaw plugin)

This skill is a shared sub-procedure invoked by the Chorus stage skills (proposal, develop, yolo) when the resolved spec mode is a usable OpenSpec — spec-driven authoring through the OpenSpec CLI:

• Activates when the resolved spec mode is a usable OpenSpec (see §1): CHORUS_SPEC_MODE=openspec or unset, and CHORUS_OPENSPEC_MODE not off, an openspec/ directory at the project root, and the openspec CLI on PATH.
• Otherwise the calling skill follows the resolved mode — spec-lite (the default when OpenSpec isn't usable) or free-form (CHORUS_SPEC_MODE=off).

See also — spec-lite (the lightweight fallback): OpenSpec (this skill) stays the default whenever it is usable. When OpenSpec is absent or disabled — or CHORUS_SPEC_MODE=lite — the mode resolves to spec-lite: a durable local .chorus/specs/<slug>/spec.md (never synced) + per-change dated folders <slug>/<YYYY-MM-DD>-<change-slug>/ of Chorus-typed docs mirrored 1:1 into Chorus via the same --arg-file transport. See the spec-lite skill.

Tool namespace: Chorus MCP tools are exposed under a chorus__ prefix on OpenClaw (e.g. chorus__chorus_pm_create_proposal). Bare names are used in prose for readability — prepend chorus__ when invoking the MCP tools directly. Document-mirror calls do NOT go through the MCP harness at all — they go through the chorus CLI (chorus mcp call, preferred) or the chorus-api.sh wrapper (fallback) (see §2 Rule 1), which talk to the Chorus MCP endpoint over HTTP using your API key, independent of the chorus__ namespacing.

§1. Resolve the spec mode — run the shipped resolver, every time (no SessionStart hook on OpenClaw)

OpenClaw difference: the Claude Code / Codex / Pi ports precompute the spec mode once (in a SessionStart hook or an extension handler) and inject a ## Spec Mode value into context. OpenClaw does none of that. You MUST resolve the mode yourself, at the moment you reach this skill from a stage skill. Do not look for an injected CHORUS_OPENSPEC_ACTIVE / ## Spec Mode value — it will not exist on OpenClaw. Resolve the whole contract (lite / openspec / off), not just "is OpenSpec active".

Never hand-roll the rule in this Markdown. The plugin ships the canonical resolver — bin/resolve-spec-mode.sh, byte-identical to the Claude Code copy and enforced so by test-resolver-drift.sh — precisely so this skill can source it instead of reproducing it. A prose copy of the rule drifts from the real one (it has); a sourced script cannot.

The rule: an explicit CHORUS_SPEC_MODE (lite/openspec/off) wins; when unset, OpenSpec is the default whenever it is usable — else spec-lite. OpenSpec is usable only when all three hold:

1. OpenSpec is not disabled (CHORUS_OPENSPEC_MODE not off; the enableOpenSpec plugin toggle not false).
2. The project root contains an openspec/ directory (i.e. someone ran openspec init here).
3. The openspec CLI is on PATH.

Both signals (2) and (3) are required because the OpenSpec authoring path needs the working directory and the CLI — having one without the other leaves the workflow unrunnable. When OpenSpec is requested but usable=false (signal 2 holds but 3 does not, say), surface a hint to the user — "OpenSpec repo detected — install with: npm i -g @fission-ai/openspec".

Resolution block (run this)

# PROJECT_ROOT is your project root — OpenClaw does not export
# CLAUDE_PROJECT_DIR, so default to $PWD. The resolver reads it.
PROJECT_ROOT="${PWD}"

# Locate the resolver the plugin ships. Covers the npm install, a global npm
# install, and a linked dev checkout; the first hit wins.
RESOLVER=""
for candidate in \
  "$PWD/node_modules/@chorus-aidlc/chorus-openclaw-plugin/bin/resolve-spec-mode.sh" \
  "$(npm root -g 2>/dev/null)/@chorus-aidlc/chorus-openclaw-plugin/bin/resolve-spec-mode.sh" \
  "$PWD/packages/openclaw-plugin/bin/resolve-spec-mode.sh"
do
  [ -f "$candidate" ] && { RESOLVER="$candidate"; break; }
done

# shellcheck source=/dev/null
. "$RESOLVER"    # sets SPEC_MODE, SPEC_REASON, SPEC_FAIL, OPENSPEC_HINT, CHORUS_OPENSPEC_ACTIVE
echo "SPEC_MODE=$SPEC_MODE CHORUS_OPENSPEC_ACTIVE=$CHORUS_OPENSPEC_ACTIVE SPEC_FAIL=${SPEC_FAIL} REASON=${SPEC_REASON}"

If RESOLVER came out empty (the loop found nothing), halt and ask the user to run /chorus spec — which resolves the same contract from the plugin's src/spec-mode.ts — and paste the result. Do not substitute your own detection.

Branch on the result:

• SPEC_FAIL non-empty (explicit CHORUS_SPEC_MODE=openspec that can't be honored — either a config conflict or OpenSpec not installed; OPENSPEC_HINT carries the install command when it is merely missing) → the caller MUST halt and surface it verbatim. Do not silently fall back.
• CHORUS_OPENSPEC_ACTIVE=1 → follow §3 (OpenSpec authoring).
• Otherwise (SPEC_MODE=lite or off, no fail) → this skill is a no-op; return to the calling skill, which follows the resolved mode: spec-lite (the default when OpenSpec isn't usable — see the spec-lite skill) or free-form (off). Do not scaffold openspec/changes/. Do not add the slug line to the proposal description.

Run this resolution whenever proposal / develop / yolo reference this skill. There is no host-injected value to read on OpenClaw — but "resolve it yourself" means run the shipped resolver, never re-derive the rule from this document's prose.

§2. ⛔ Two non-negotiable rules

Both are enforced at review time. Both have caused incidents in past releases.

Rule 1 — Fill content from the file (CLI preferred, bash-wrapper fallback); never re-type document content from agent output

Document/draft mirror calls (chorus_pm_add_document_draft, chorus_pm_update_document_draft, chorus_pm_update_document) MUST fill the content field from the local file's bytes, never from a hand-typed body. Calling these tools directly from the agent's MCP harness with a hand-typed content field is a protocol violation for OpenSpec mode and will fail review. Use whichever transport is available, preferred first:

• Primary — the chorus CLI: chorus mcp call <tool_name> '<json-without-content>' --arg-file content=<file>. --arg-file content=<path> reads the file's raw bytes and injects them as the JSON content string, byte-exact — the CLI's built-in replacement for json_encode_file, so no helper is needed. chorus mcp call reads the same CHORUS_URL / CHORUS_API_KEY from the environment (availability note below). See §3.6. Requires chorus >= 0.17.0 (the chorus mcp subcommand was added then; an older CLI errors with "unknown command"); on any version or unknown-command failure, upgrade with npm install -g @chorus-aidlc/chorus.
• Fallback — the chorus-api.sh wrapper, when chorus is not on PATH: build $PAYLOAD with the json_encode_file helper and call chorus-api.sh mcp-tool <tool_name> "$PAYLOAD". Defined in the §3.6 fallback block.

New to the chorus CLI? See the chorus-cli skill for install, configuring agents (chorus agents add|remove|list), the connection env vars, and chorus mcp basics.

Acting identity — which agent the call acts as. chorus mcp call resolves the agent from, in order: CHORUS_AGENT_PROFILE (a name or UUID) → CHORUS_URL + CHORUS_API_KEY in the environment → the single agent configured in ~/.chorus/daemon.json. A daemon-woken session already has CHORUS_AGENT_PROFILE set. If a mirror call fails with Multiple agents … specify --agent (several agents configured and no profile/creds in the env), pass your own identity explicitly: chorus mcp call <tool> … --agent <your-agentUuid> — your UUID is in your chorus_checkin result, and chorus agents lists every configured name/UUID.

Reasons (they apply to both paths):

1. Token cost. Re-typing a multi-thousand-line markdown body through the LLM burns input + output tokens for every draft. Both the CLI's --arg-file and the fallback's json_encode_file stream the file's bytes into the JSON string — content never enters LLM context. A typical 3-doc proposal mirror costs roughly zero content-tokens this way; via direct MCP with a re-typed body it routinely costs 20k+.
2. Byte-equality. A file-fill path (CLI --arg-file, or the fallback's jq -Rs '.') is a byte-faithful encoder: backslashes, quotes, newlines, code-fence content, zero-width chars all survive. LLM re-emission has a non-zero failure rate on long markdown — table alignment drifts, fence escapes get "fixed", long URLs wrap. The byte-equality guarantee (modulo trailing \n) holds only on a file-fill path, never on LLM re-emission.
3. Single source of truth. With a file-fill mirror, the local openspec/changes/<slug>/*.md is authoritative and Chorus is a mirror. With agent re-typing, authority splits between local file and whatever the LLM happened to output — a future diff cannot tell which one is correct.

chorus / chorus-api.sh availability on OpenClaw. The document-mirror transport is either the chorus CLI (preferred) or the chorus-api.sh wrapper (fallback). The wrapper must be reachable as chorus-api.sh on PATH (the Chorus standalone skill bundle ships it; if you installed via that bundle it is already on PATH). If neither chorus nor chorus-api.sh is on PATH in your OpenClaw environment, do one of:

• call the wrapper by its absolute path (e.g. "$HOME/.chorus/bin/chorus-api.sh" mcp-tool ...), or
• reproduce its single behavior inline — POST a JSON-RPC tools/call for <tool_name> with arguments $PAYLOAD to "$CHORUS_URL/api/mcp" with header Authorization: Bearer $CHORUS_API_KEY using curl, capturing the raw body for the §6 halt-on-error check.

Both chorus mcp call and the wrapper read CHORUS_URL and CHORUS_API_KEY from the environment (same values as your plugin config chorusUrl / apiKey). Export them before the first call if they are not already set. What you must NOT do is re-type the document body through the model — the file-fill rule stands regardless of which transport you use.

Rule 2 — Halt on error via chorus_check_response

Every wrapper call must check three signals: wrapper exit code, "error": in body, empty body. Bare RC=$? is insufficient — the wrapper exits 0 on HTTP 401 (auth failure) with empty body, so a single-signal check silently misses the most common runtime failure. See §6 for the helper definition.

§3. OpenSpec mode authoring

3.1 Pick a slug

openspec/changes/<slug>/ is the local change folder. The slug must be:

• kebab-case (add-export-csv, not addExportCsv or add_export_csv),
• derived from the source Idea title,
• unique within openspec/changes/.

Record it for later steps:

SLUG="add-export-csv"

3.2 Scaffold the change folder

openspec new change "$SLUG" --description "<one-line idea summary>"

This creates openspec/changes/$SLUG/ with README.md and .openspec.yaml. Then author by hand:

Use openspec instructions <artifact> --change "$SLUG" (artifacts: proposal, specs, design, tasks) for templates.

3.3 Spec file shape (verified against openspec instructions specs)

A delta spec lists one or more block headers — ## ADDED Requirements, ## MODIFIED Requirements, ## REMOVED Requirements, ## RENAMED Requirements — and within each, ### Requirement: entries. Mix freely in the same file; only include the blocks you actually need.

## ADDED Requirements

Append a brand-new Requirement to the long-term spec.

## ADDED Requirements

### Requirement: <name>
<requirement text — use SHALL / MUST for normative behavior>

#### Scenario: <name>
- **WHEN** <condition>
- **THEN** <expected outcome>

## MODIFIED Requirements

Whole-block replacement, not merge. Whatever you write here completely replaces the existing same-named Requirement in the long-term spec — title, description, and all scenarios. Half-writing it deletes the rest.

## MODIFIED Requirements

### Requirement: <existing name>
<full updated requirement text>

#### Scenario: <name>
- **WHEN** <condition>
- **THEN** <expected outcome>

#### Scenario: <other name>
- **WHEN** <condition>
- **THEN** <expected outcome>

Always include every scenario you want the post-archive spec to have, even ones that were already present and unchanged.

## REMOVED Requirements

Delete a Requirement from the long-term spec. The block under the heading is just the requirement name(s) you're removing — no scenarios needed.

## REMOVED Requirements

### Requirement: <existing name>

## RENAMED Requirements

Rename a Requirement's title. Body and scenarios are preserved as-is in the long-term spec; use MODIFIED instead if you need to change anything besides the title.

## RENAMED Requirements

### Requirement: <old name> -> <new name>

Critical formatting rules (verified):

• Scenarios MUST use exactly 4 hashtags (#### Scenario:). 3 hashtags or a bullet list silently fail validation.
• Every ### Requirement: under ADDED or MODIFIED MUST have at least one #### Scenario:.
• MODIFIED blocks MUST include the full updated content — they overwrite, not patch.
• Use SHALL / MUST for normative requirements; avoid should / may.
• The merge into openspec/specs/<capability>/spec.md happens at openspec archive time (§3.9), not at proposal time. While the proposal is in flight, Chorus only sees the delta file as one spec Document — there is no half-merged state for the skill to reason about.

Optional:

openspec validate "$SLUG"

3.4 Filling the content field byte-exact

The document content must be inserted byte-for-byte from the local file — never re-typed by the LLM. Two mechanisms, preferred first:

• Primary — chorus mcp call … --arg-file content=<path> (§3.6). The CLI reads the file's raw bytes and injects them as the JSON content string. This is the byte-faithful replacement for json_encode_file, so on the CLI path no helper is needed — pass the base JSON without a content field and let --arg-file fill it.
• Fallback — json_encode_file (defined in the §3.6 fallback block, used only when chorus is not on PATH). With jq available it streams the file through jq -Rs '.'; the pure-shell branch matches chorus-api.sh's own escaping when jq is missing.

Round-trip: the Chorus backend appends a single \n to draft content on write, so server content is byte-equal modulo a trailing newline. Reviewers diffing local file vs server should ignore that one byte.

3.5 Create the proposal container with the slug provenance line

Use the regular chorus_pm_create_proposal MCP tool (no wrapper required for this single call — the description is short, the LLM-emitted version is fine). The description must carry exactly one line:

OpenSpec change slug: <slug>

• on its own line (no other text on that line),
• literal prefix OpenSpec change slug: (capital O, capital S, single space after colon),
• no trailing punctuation,
• value matches the slug passed to openspec new change.

This line is machine-grep-able by future runs of this skill and by the §3.9 archive trigger.