@codedrifters/configulator 0.0.262 → 0.0.264

package/lib/index.mjs

@@ -5187,6 +5187,86 @@ var meetingAnalystSubAgent = {
     "5. **Bi-directional traceability.** Every document and issue created by",
     "   this pipeline must link back to the meeting source, and the meeting",
     "   source documents must link forward to everything created from them.",
+    "6. **Respect meeting type and areas.** Read the `meeting_type` and",
+    "   `areas` values from the meeting frontmatter before starting Phase 1.",
+    "   Apply the type-specific rules from the **Meeting type handling**",
+    "   table (phases 1\u20132) and the area-filtered routing rules from the",
+    "   **Areas filtering** table (phase 4).",
+    "",
+    "---",
+    "",
+    "## Meeting type handling",
+    "",
+    "Meeting frontmatter is expected to carry a `meeting_type` field whose",
+    "value is one of the project's concrete meeting-type `id`s (declared",
+    "via `AgentConfigOptions.meetings.meetingTypes`). Each concrete `id`",
+    "maps to one of six generic **meeting kinds** that drive type-specific",
+    "extraction rules:",
+    "",
+    "| Kind | Primary outputs | Phase-1/2 rules |",
+    "|------|-----------------|-----------------|",
+    "| `planning` | Sprint/plan updates, task issues, requirement drafts | Extract every task assignment. Phase 4 updates the active sprint-plan doc directly and creates one issue per assigned task. |",
+    "| `review` | Retro notes, follow-up issues, status updates, requirement revisions | Mark completed tasks in the active sprint-plan doc. Capture retrospective learnings inside the meeting notes \u2014 do not spawn a separate retro document. |",
+    "| `brainstorm` | Future-feature candidates, research topics, very few Firm decisions | Lower the bar for Open Questions. Do **not** create requirement or ADR issues from brainstorm output unless the item is explicitly marked **Firm**. Prefer research/discovery issues for promising ideas. |",
+    "| `standup` | Action items, blockers, status updates | Do not produce requirement, ADR, or BDR drafts. Phase 3 (Draft) is almost always skipped. Focus on action-item issues and sprint-plan status updates. |",
+    "| `external` | Customer/competitive intel, people/company profiles, lead capture | Default every attendee outside the team into **People of Interest**. Default every organization mentioned into **Companies of Interest** (signal threshold still applies). Capture stated customer pain points as candidate **BR** (business requirements), not FR. |",
+    "| `other` | General meeting extraction | Apply the default workflow with no type-specific overrides. |",
+    "",
+    "**How to resolve the kind:**",
+    "",
+    "1. Read `meeting_type` from the frontmatter.",
+    "2. Look up the matching `id` in the project's **Recognized meeting",
+    "   types** table (rendered below when `meetingTypes` is declared).",
+    "3. Use the `Kind` column value. If the table is absent, the `id` is",
+    "   unknown, or the `Kind` column is `\u2014`, fall back to `other`.",
+    "4. Apply the rules in the row above **in addition to** the normal",
+    "   phase workflow \u2014 never as a replacement.",
+    "",
+    "If `meeting_type` is missing, invalid, or the declared `id` is not in",
+    "the project's table, note it in the extraction's Open Questions and",
+    "proceed with the default workflow (treat as `other`).",
+    "",
+    "## Areas filtering",
+    "",
+    "Meeting frontmatter may carry an `areas: [...]` list whose values",
+    "match `id`s from the project's **Area \u2192 doc-root mapping** table",
+    "(declared via `AgentConfigOptions.meetings.meetingAreas`). The list",
+    "scopes phase-4 direct edits and cross-references to the doc-root",
+    "sub-folders registered for those areas. A given area's effective",
+    "path resolves as `<docsRoot>/<docRoot>` \u2014 `docsRoot` comes from",
+    "`AgentPathsConfig.docsRoot`; `docRoot` is the sub-folder shown in",
+    "the **Area \u2192 doc-root mapping** table.",
+    "",
+    "**Gating contract (phase 4 only):**",
+    "",
+    "- **Scoped in:** direct edits and cross-reference stubs under the",
+    "  `docRoot` of every `id` present in the meeting's `areas` list.",
+    "- **Scoped out:** direct edits to any doc-root sub-folder that is",
+    "  declared in the project's area map but whose `id` is **not** in",
+    "  the meeting's `areas`. For those out-of-scope areas, create",
+    "  follow-up issues instead of editing the doc directly.",
+    "- **Not gated:** issue creation (requirements, ADRs, profiles,",
+    "  research) is always in scope regardless of areas.",
+    "",
+    "**Resolution rules:**",
+    "",
+    "1. If the meeting's `areas` frontmatter is missing or empty **and**",
+    "   the project has declared a non-empty `meetingAreas` map, no",
+    "   doc-root sub-folder is gated in: direct edits to any declared",
+    "   `docRoot` sub-folder are out-of-scope and become follow-up issues",
+    "   per step 6. Only the default meetings root (`meetingsRoot`)",
+    "   remains editable.",
+    "2. If `areas` lists an `id` that is not in the project's area map,",
+    "   note it in the extraction's Open Questions and treat the unknown",
+    "   `id` as absent. Do **not** invent a doc-root path for it.",
+    "3. Out-of-scope direct edits become `type:docs` follow-up issues on",
+    "   phase 4 step 6, with the area noted in the issue body.",
+    "",
+    "When the project has **not** declared any `meetingAreas` at all,",
+    "this section applies degenerately \u2014 no areas are gated, no",
+    "follow-up issues are synthesized for missing areas, and phase-4",
+    "direct edits follow the default workflow (step 6 edits any doc",
+    "flagged as Firm, unrestricted by `docRoot`).",
     "",
     "---",
     "",
@@ -5244,7 +5324,13 @@ var meetingAnalystSubAgent = {
     "",
     "### Steps",
     "",
-    "1. Read the transcript file specified in the issue body.",
+    "1. Read the transcript file specified in the issue body. Parse the",
+    "   meeting-note frontmatter first, capturing `meeting_type` and",
+    "   `areas`. Resolve `meeting_type` to a kind using the **Meeting type",
+    "   handling** table; note the resolved kind in the extraction",
+    "   frontmatter. If `meeting_type` is missing or not in the project's",
+    "   Recognized meeting types table, treat it as `other` and flag it",
+    "   in Open Questions.",
     "2. Identify and categorize content into these buckets:",
     "",
     "   | Bucket | What to look for |",
@@ -5270,14 +5356,32 @@ var meetingAnalystSubAgent = {
     "   - Companies of Interest (with type and context)",
     "   - Strategic / Product Direction",
     "",
-    "4. **Create downstream phase issues** using `gh issue create`:",
+    "   Carry `meeting_type`, the resolved `meeting_kind`, and `areas`",
+    "   into the extraction frontmatter so downstream phases can gate",
+    "   behaviour without re-parsing the meeting note.",
+    "",
+    "4. **Apply type-specific rules.** Consult the **Meeting type handling**",
+    "   table for the resolved kind and apply its phase-1 rules in",
+    "   addition to the default extraction above. Examples:",
+    "   - `standup`: skip requirement, ADR, and BDR extraction; focus on",
+    "     action items and status.",
+    "   - `brainstorm`: emphasize Open Questions and Future Features.",
+    "     Only mark decisions as Firm when the transcript is unambiguous.",
+    "   - `external`: default attendees outside the team into People of",
+    "     Interest; default organizations mentioned into Companies of",
+    "     Interest (signal threshold still applies); capture customer",
+    "     pain points as candidate BR (not FR).",
+    "",
+    "5. **Create downstream phase issues** using `gh issue create`:",
     "   - Always create a `meeting:notes` issue (blocked on this extract issue)",
     "   - If requirements OR decisions/ADRs identified, create a `meeting:draft` issue",
-    "     (blocked on the notes issue)",
+    "     (blocked on the notes issue). For kind `standup`, skip the",
+    "     `meeting:draft` issue \u2014 standup meetings almost never produce",
+    "     formal drafts.",
     "   - Always create a `meeting:link` issue \u2014 blocked on the draft issue if one",
     "     was created, otherwise blocked on the notes issue",
     "",
-    "5. Commit, push, and close the extract issue.",
+    "6. Commit, push, and close the extract issue.",
     "",
     "---",
     "",
@@ -5334,18 +5438,38 @@ var meetingAnalystSubAgent = {
     "### Steps",
     "",
     "1. Read the drafts from Phase 3 (if they exist) and the extraction from Phase 1.",
+    "   Re-read the extraction frontmatter for `meeting_kind` and `areas`,",
+    "   then build the area-gate: the set of `docRoot` sub-folders that",
+    "   are in scope for direct edits on this meeting. Apply the rules in",
+    "   the **Areas filtering** section above.",
     "2. Create requirement issues using `gh issue create` with appropriate labels.",
     "   Include a `## Traceability` section in each issue body linking back to",
-    "   the source meeting and extraction file.",
+    "   the source meeting and extraction file. Issue creation is **not**",
+    "   gated by areas.",
     "3. Create ADR issues if technology decisions need formal records.",
-    "   Include a `## Traceability` section in each.",
+    "   Include a `## Traceability` section in each. Not gated by areas.",
     "4. Create action item issues for tasks that are not document-related.",
-    "   Include a `## Traceability` section in each.",
-    "5. Cross-reference the meeting in any existing documents that were discussed.",
-    "6. Apply direct product/strategy doc updates for items flagged as **Firm**",
-    "   confidence in the extraction.",
+    "   Include a `## Traceability` section in each. Not gated by areas.",
+    "5. Cross-reference the meeting in any existing documents that were",
+    "   discussed **and that live under an in-scope `docRoot`**. For",
+    "   discussed documents under an out-of-scope `docRoot`, open a",
+    "   follow-up `type:docs` issue instead of editing the doc directly.",
+    "6. Apply direct product/strategy doc updates for items flagged as",
+    "   **Firm** confidence. When the project declares `meetingAreas`,",
+    "   gate these edits: only apply them when the target path lies",
+    "   under an in-scope `docRoot`, and for Firm items targeting an",
+    "   out-of-scope `docRoot` open a follow-up `type:docs` issue",
+    "   noting the area and the target path. When the project has not",
+    "   declared any `meetingAreas`, apply Firm updates unrestricted",
+    "   (the degenerate case above). For the kind `planning`, also",
+    "   update the active sprint-plan doc directly (when the",
+    "   sprint-plan doc's area is in-scope, or always when no",
+    "   `meetingAreas` are declared); for the kind `review`, mark",
+    "   completed tasks in that same doc.",
     "7. **Update the extraction file** with a `## Downstream Artifacts` section",
-    "   listing every issue and document created from this meeting.",
+    "   listing every issue and document created from this meeting. Note",
+    "   any items that were deferred to follow-up issues because of the",
+    "   area gate.",
     "8. **Update the meeting notes** with a `## Related Issues` section listing",
     "   all issues created from this meeting.",
     "9. Comment on the parent extract issue with a summary linking to all",
@@ -5429,7 +5553,17 @@ var meetingAnalysisBundle = {
         "Use `/process-meeting <path>` to process a meeting transcript through a",
         "4-phase pipeline (extract \u2192 notes \u2192 draft \u2192 link). Each phase runs as a",
         "separate agent session tracked by a GitHub issue with a `meeting:*` label.",
-        "See the `meeting-analyst` agent definition for full workflow details."
+        "See the `meeting-analyst` agent definition for full workflow details.",
+        "",
+        "Meeting notes may declare a `meeting_type` (one of the project's",
+        "recognized types) and an `areas: [...]` list. The `meeting_type`",
+        "resolves to a generic kind \u2014 `planning` / `review` / `brainstorm` /",
+        "`standup` / `external` / `other` \u2014 that drives type-specific",
+        "extraction rules in phases 1\u20132. The `areas` list scopes phase-4",
+        "direct edits to the doc-root sub-folders declared in the project's",
+        "**Area \u2192 doc-root mapping** table. When both fields are absent",
+        "the pipeline falls back to the default workflow (kind `other`,",
+        "no area gating)."
       ].join("\n"),
       platforms: {
         cursor: { exclude: true }
@@ -9199,7 +9333,7 @@ var REQ_WRITE_ISSUE_SCHEMA_SECTION = [
 // src/agent/bundles/requirements-analyst.ts
 var requirementsAnalystSubAgent = {
   name: "requirements-analyst",
-  description: "Discovers requirement gaps from BCM model docs, competitive analysis, product docs, and meeting extracts. Produces scan reports and proposals for the downstream requirements-writer agent. Runs through a 3-phase pipeline (scan \u2192 draft \u2192 trace), one phase per session, tracked by req:* GitHub issue labels.",
+  description: "Discovers requirement gaps from BCM model docs, competitive analysis, product docs, and meeting extracts. Produces scan reports, proposals, and req:write issues for the downstream requirements-writer agent. Runs through a 2-phase pipeline (scan \u2192 draft-trace), one phase per session, tracked by req:* GitHub issue labels.",
   model: AGENT_MODEL.POWERFUL,
   maxTurns: 80,
   platforms: { cursor: { exclude: true } },
@@ -9240,35 +9374,39 @@ var requirementsAnalystSubAgent = {
     "",
     "## State Machine Overview",
     "",
-    "Requirements synthesis flows through **3 phases**:",
+    "Requirements synthesis flows through **2 phases**:",
     "",
     "```",
-    "\u250C\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510     \u250C\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510     \u250C\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510",
-    "\u2502  1. SCAN     \u2502\u2500\u2500\u2500\u2500\u25B6\u2502  2. DRAFT    \u2502\u2500\u2500\u2500\u2500\u25B6\u2502  3. TRACE    \u2502",
-    "\u2502  Read docs,  \u2502     \u2502  Write gap   \u2502     \u2502  Create GH   \u2502",
-    "\u2502  identify    \u2502     \u2502  report with \u2502     \u2502  issues and  \u2502",
-    "\u2502  gaps, check \u2502     \u2502  requirement \u2502     \u2502  update src  \u2502",
-    "\u2502  for dupes   \u2502     \u2502  proposals   \u2502     \u2502  docs with   \u2502",
-    "\u2502              \u2502     \u2502              \u2502     \u2502  traceability\u2502",
-    "\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518     \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518     \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518",
+    "\u250C\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510     \u250C\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510",
+    "\u2502  1. SCAN     \u2502\u2500\u2500\u2500\u2500\u25B6\u2502  2. DRAFT-TRACE  \u2502",
+    "\u2502  Read docs,  \u2502     \u2502  Write proposals,\u2502",
+    "\u2502  identify    \u2502     \u2502  create req:write\u2502",
+    "\u2502  gaps, check \u2502     \u2502  issues, update  \u2502",
+    "\u2502  for dupes   \u2502     \u2502  source docs     \u2502",
+    "\u2502              \u2502     \u2502  with traceability\u2502",
+    "\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518     \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518",
     "```",
     "",
+    "This pipeline matches the `scan \u2192 draft-trace \u2192 write` pattern",
+    "used by the sibling openhi project. Draft and Trace were previously",
+    "separate phases; they were collapsed because the proposals file Draft",
+    "wrote was only ever consumed by Trace (no human review, no async work",
+    "between the two), so the phase boundary added latency without value.",
+    "See ADR-009 for the full reasoning.",
+    "",
     "**Issue labels encode the phase:**",
     "",
     "| Label | Phase | Session work |",
     "|-------|-------|-------------|",
     "| `req:scan` | 1. Scan | Read source docs, identify potential requirement gaps, check against existing requirements and open issues, write deduplicated scan report |",
-    "| `req:draft` | 2. Draft | Write gap report with proposed requirements |",
-    "| `req:trace` | 3. Trace | Create GitHub issues for each proposed requirement and update source documents with traceability notes |",
+    "| `req:draft-trace` | 2. Draft & Trace | Write proposals, create `req:write` GitHub issues for each proposal, and update source documents with traceability notes \u2014 all in a single session |",
     "",
     "All issues also carry `type:requirement` and a `status:*` label.",
     "",
-    "**Issue count per scan cycle:** 1 scan + 1 draft + 1 trace = **3 sessions**.",
+    "**Issue count per scan cycle:** 1 scan + 1 draft-trace = **2 sessions**.",
     "",
     "**Shortened paths:**",
-    "- No gaps found after scan \u2192 skip draft and trace \u2192 **1 session**",
-    "- No source docs need traceability updates \u2192 still **3 sessions** (trace",
-    "  handles both issue creation and doc updates)",
+    "- No gaps found after scan \u2192 skip draft-trace \u2192 **1 session**",
     "",
     "---",
     "",
@@ -9297,7 +9435,7 @@ var requirementsAnalystSubAgent = {
     "Run this loop exactly once per session. Never start a second issue.",
     "",
     "1. Claim one open `type:requirement` issue using phase priority:",
-    "   `req:scan` > `req:draft` > `req:trace`.",
+    "   `req:scan` > `req:draft-trace`.",
     "2. Transition `status:ready` \u2192 `status:in-progress` and create the branch",
     "   per your project's branch-naming convention.",
     "3. Execute the phase handler that matches the issue's `req:*` label.",
@@ -9398,7 +9536,7 @@ var requirementsAnalystSubAgent = {
     "   ```",
     "",
     "7. **Create downstream issues based on findings:**",
-    "   - If any new gaps were identified \u2192 create `req:draft` issue",
+    "   - If any new gaps were identified \u2192 create `req:draft-trace` issue",
     "     (blocked on this issue via `Depends on: #N`).",
     "   - If **no gaps** were found \u2192 stop (no further phases needed). Comment",
     "     on the issue noting that no gaps were identified, and proceed directly",
@@ -9409,12 +9547,54 @@ var requirementsAnalystSubAgent = {
     "",
     "---",
     "",
-    "## Phase 2: Draft (`req:draft`)",
+    "## Phase 2: Draft & Trace (`req:draft-trace`)",
+    "",
+    "**Goal:** Expand each identified gap into a requirement proposal, create",
+    "a `req:write` GitHub issue for each proposal so the downstream",
+    "`requirements-writer` bundle picks it up, and backfill source-document",
+    "traceability notes \u2014 all in a single session.",
+    "",
+    "**Budget:** No web searches. Reading + writing proposals + issue creation",
+    "+ minor traceability edits to source documents.",
+    "",
+    "Draft and Trace were previously separate phases. They were collapsed",
+    "because the proposals file written by Draft was only ever consumed by",
+    "Trace \u2014 no human review, no async work, no CI validation sat between",
+    "them. See ADR-009 for the decision record.",
+    "",
+    "### `req:write` issue schema",
+    "",
+    "The `req:write` issues this phase creates are picked up by the",
+    "downstream `requirements-writer` agent, which parses a strict",
+    "schema on intake. The authoritative schema is defined in the",
+    "`requirements-writer` sub-agent prompt under **The `req:write`",
+    "Issue Schema** \u2014 every `req:write` issue this phase opens must",
+    "conform. The same schema is embedded in the",
+    "`requirements-reviewer` follow-up generator so the three",
+    "producers never drift apart.",
+    "",
+    ...REQ_WRITE_ISSUE_SCHEMA_SECTION,
     "",
-    "**Goal:** Expand each identified gap into a requirement proposal with enough",
-    "detail for the requirements-writer to produce a full document.",
+    "**Deriving the three required fields from the proposal.** For",
+    "every `req:write` issue this phase opens, the three fields come",
+    "directly from the proposal entry written earlier in the same",
+    "session:",
     "",
-    "**Budget:** No web searches. Reading + writing.",
+    "- **Category** \u2014 the proposal's `**Category:**` line.",
+    "- **Tier** \u2014 the proposal's `**Tier:**` line.",
+    "- **Output Path** \u2014",
+    "  `<REQUIREMENTS_ROOT>/<category-dir>/<PREFIX>-<NNN>-<slug>.md`,",
+    "  using the sequence number determined in the proposal step below.",
+    "",
+    "**Validation step \u2014 required before opening the issue.** Before",
+    "calling `gh issue create`, verify all three required fields are",
+    "populated and internally consistent (Category matches title",
+    "prefix, Tier matches `tier:*` label, Output Path filename starts",
+    "with the Category prefix). If any field cannot be derived, open",
+    "the issue with `status:needs-attention` (not `status:ready`) and",
+    "include a `Missing: <field> \u2014 <one-line reason>` line so a human",
+    "triaging the issue only has to supply the remaining value(s)",
+    "before flipping the label to `status:ready`.",
     "",
     "### Steps",
     "",
@@ -9465,61 +9645,7 @@ var requirementsAnalystSubAgent = {
     "   directory under `<REQUIREMENTS_ROOT>/<category>/` to find the next",
     "   available `NNN` for each proposed requirement.",
     "",
-    "5. **Create the `req:trace` issue** blocked on this draft issue.",
-    "",
-    "6. **Commit and push.**",
-    "",
-    "---",
-    "",
-    "## Phase 3: Trace (`req:trace`)",
-    "",
-    "**Goal:** Create GitHub issues for each proposed requirement and update",
-    "source documents with traceability notes indicating that requirement issues",
-    "were created.",
-    "",
-    "**Budget:** No web searches. Issue creation + minor edits to source",
-    "documents.",
-    "",
-    "### `req:write` issue schema",
-    "",
-    "The `req:write` issues this phase creates are picked up by the",
-    "downstream `requirements-writer` agent, which parses a strict",
-    "schema on intake. The authoritative schema is defined in the",
-    "`requirements-writer` sub-agent prompt under **The `req:write`",
-    "Issue Schema** \u2014 every `req:write` issue this phase opens must",
-    "conform. The same schema is embedded in the",
-    "`requirements-reviewer` follow-up generator so the three",
-    "producers never drift apart.",
-    "",
-    ...REQ_WRITE_ISSUE_SCHEMA_SECTION,
-    "",
-    "**Deriving the three required fields from the proposal.** For",
-    "every `req:write` issue this phase opens, the three fields come",
-    "directly from the Phase 2 proposal entry:",
-    "",
-    "- **Category** \u2014 the proposal's `**Category:**` line.",
-    "- **Tier** \u2014 the proposal's `**Tier:**` line (added in Phase 2).",
-    "  If the proposal omitted the Tier line (older proposals may),",
-    "  route the issue through `Missing:` rather than guessing.",
-    "- **Output Path** \u2014",
-    "  `<REQUIREMENTS_ROOT>/<category-dir>/<PREFIX>-<NNN>-<slug>.md`,",
-    "  using the sequence number determined in Phase 2 step 4.",
-    "",
-    "**Validation step \u2014 required before opening the issue.** Before",
-    "calling `gh issue create`, verify all three required fields are",
-    "populated and internally consistent (Category matches title",
-    "prefix, Tier matches `tier:*` label, Output Path filename starts",
-    "with the Category prefix). If any field cannot be derived, open",
-    "the issue with `status:needs-attention` (not `status:ready`) and",
-    "include a `Missing: <field> \u2014 <one-line reason>` line so a human",
-    "triaging the issue only has to supply the remaining value(s)",
-    "before flipping the label to `status:ready`.",
-    "",
-    "### Steps",
-    "",
-    "1. **Read the proposals** from Phase 2.",
-    "",
-    "2. **Create requirement issues.** For each proposal:",
+    "5. **Create requirement issues.** For each proposal:",
     "",
     "   All `type:requirement` issues default to `priority:medium` (override",
     "   only if the proposal's priority was explicitly High or Low). Each",
@@ -9564,7 +9690,7 @@ var requirementsAnalystSubAgent = {
     "   whichever of the three fields **could** be derived so a human",
     "   triaging the issue has the minimum possible cleanup.",
     "",
-    "3. **Update source documents.** In each BCM model doc or competitive",
+    "6. **Update source documents.** In each BCM model doc or competitive",
     "   analysis that was scanned, add a note in the project-relevance /",
     "   strategic-implications section (whichever heading the source doc uses)",
     "   indicating that a requirement issue was created:",
@@ -9574,10 +9700,10 @@ var requirementsAnalystSubAgent = {
     "     (issue #<N>)",
     "   ```",
     "",
-    "4. **Comment on the scan issue** with a summary of all issues created and",
+    "7. **Comment on the scan issue** with a summary of all issues created and",
     "   all docs updated.",
     "",
-    "5. **Commit and push.**",
+    "8. **Commit and push.**",
     "",
     "---",
     "",
@@ -9655,39 +9781,41 @@ var scanRequirementsSkill = {
     "1. Create a `req:scan` issue with `type:requirement`, `priority:medium`,",
     "   and `status:ready`. Body must list the files to read and the scan scope.",
     "2. Execute Phase 1 (Scan) of the requirements-analyst agent.",
-    "3. If gaps are found, a `req:draft` issue is created automatically.",
+    "3. If gaps are found, a `req:draft-trace` issue is created automatically.",
     "",
     "## Output",
     "",
     "- A `req-scan-<scope>-<YYYY-MM-DD>.md` file under the project's research",
     "  requirements directory.",
-    "- A `req:draft` issue if any gaps were identified."
+    "- A `req:draft-trace` issue if any gaps were identified."
   ].join("\n")
 };
 var requirementsAnalystBundle = {
   name: "requirements-analyst",
-  description: "Requirements gap-discovery agent bundle for BCM-driven projects. 3-phase pipeline (scan, draft, trace) with req:* phase labels.",
+  description: "Requirements gap-discovery agent bundle for BCM-driven projects. 2-phase pipeline (scan, draft-trace) with req:* phase labels.",
   appliesWhen: () => true,
   rules: [
     {
       name: "requirements-analyst-workflow",
-      description: "Describes the 3-phase requirements gap-discovery pipeline, the req:* label taxonomy, and the boundary with the downstream requirements-writer agent.",
+      description: "Describes the 2-phase requirements gap-discovery pipeline, the req:* label taxonomy, and the boundary with the downstream requirements-writer agent.",
       scope: AGENT_RULE_SCOPE.ALWAYS,
       content: [
         "# Requirements Analyst Workflow",
         "",
         "Use `/scan-requirements <scope>` to kick off a requirements gap",
-        "discovery cycle. The pipeline runs in 3 phases \u2014 scan, draft, trace \u2014",
-        "each tracked by its own GitHub issue labeled `req:scan`, `req:draft`,",
-        "or `req:trace`. All issues carry `type:requirement`.",
-        "",
-        "The requirements-analyst *discovers gaps and drafts proposals*; it",
-        "does **not** write final requirement documents. Writing is the job",
-        "of the downstream `requirements-writer` agent (a separate bundle).",
+        "discovery cycle. The pipeline runs in 2 phases \u2014 scan and",
+        "draft-trace \u2014 each tracked by its own GitHub issue labeled",
+        "`req:scan` or `req:draft-trace`. All issues carry",
+        "`type:requirement`.",
+        "",
+        "The requirements-analyst *discovers gaps, drafts proposals, and",
+        "opens `req:write` issues for the downstream writer*; it does",
+        "**not** write final requirement documents. Writing is the job of",
+        "the downstream `requirements-writer` agent (a separate bundle).",
         "Keep that boundary clean: proposals land under the research",
         "requirements directory, not under the authoritative requirements",
-        "tree. The trace phase tags new issues with `req:write` so the",
-        "writer bundle picks them up automatically.",
+        "tree. The draft-trace phase tags new issues with `req:write` so",
+        "the writer bundle picks them up automatically.",
         "",
         "See the `requirements-analyst` agent definition for full workflow",
         "details and phase-by-phase instructions."
@@ -9712,14 +9840,9 @@ var requirementsAnalystBundle = {
       description: "Phase 1: scan source docs for requirement gaps and deduplicate"
     },
     {
-      name: "req:draft",
+      name: "req:draft-trace",
       color: "BFDADC",
-      description: "Phase 2: draft requirement proposals for the requirements-writer"
-    },
-    {
-      name: "req:trace",
-      color: "D4C5F9",
-      description: "Phase 3: create requirement issues and backfill traceability on source docs"
+      description: "Phase 2: draft proposals, create req:write issues, and backfill source-doc traceability"
     }
   ]
 };
@@ -11603,8 +11726,8 @@ var requirementsWriterSubAgent = {
     "This agent produces **requirement documents only** \u2014 capability",
     "models are written by the `bcm-writer` agent and requirement-gap",
     "discovery is the responsibility of the `requirements-analyst` agent.",
-    "Keep this boundary clean: never open `req:scan`, `req:draft`, or",
-    "`bcm:*` issues from this pipeline.",
+    "Keep this boundary clean: never open `req:scan`,",
+    "`req:draft-trace`, or `bcm:*` issues from this pipeline.",
     "",
     "Follow your project's shared agent conventions (`AGENTS.md`,",
     "`CLAUDE.md`, or equivalent) for all commit, branch, and PR rules.",
@@ -12138,7 +12261,7 @@ var requirementsWriterSubAgent = {
     "specialized upstream/downstream agents.",
     "",
     "**Do NOT create:**",
-    "- `req:scan`, `req:draft`, or `req:trace` issues \u2014 those belong to",
+    "- `req:scan` or `req:draft-trace` issues \u2014 those belong to",
     "  the `requirements-analyst` bundle",
     "- `bcm:*` issues \u2014 those belong to the `bcm-writer` bundle",
     "- `people:*`, `company:*`, `software:*`, `research:*`, or",
@@ -12186,8 +12309,8 @@ var requirementsWriterSubAgent = {
     "  exists as a file but is missing from its category index is",
     "  invisible to anyone browsing the documentation tree.",
     "- **Write requirements, not capability models or gap reports.**",
-    "  Never open `req:scan`, `req:draft`, `req:trace`, or `bcm:*` issues",
-    "  from this pipeline."
+    "  Never open `req:scan`, `req:draft-trace`, or `bcm:*` issues from",
+    "  this pipeline."
   ].join("\n")
 };
 var writeRequirementSkill = {
@@ -12282,8 +12405,8 @@ var requirementsWriterBundle = {
         "The pipeline produces **requirement documents only** \u2014 capability",
         "models are written by the `bcm-writer` agent and gap discovery is",
         "the responsibility of the `requirements-analyst` agent. The",
-        "writer never opens `req:scan`, `req:draft`, `req:trace`, or",
-        "`bcm:*` issues.",
+        "writer never opens `req:scan`, `req:draft-trace`, or `bcm:*`",
+        "issues.",
         "",
         "Documents follow the 11-category taxonomy (BR, FR, NFR, TR, ADR,",
         "SEC, DR, INT, OPS, UX, MT) and the four-tier classification",
@@ -12389,8 +12512,8 @@ var requirementsReviewerSubAgent = {
     "`requirements-writer` bundle (requirement docs) and `bcm-writer`",
     "bundle (capability models). Gap discovery belongs to the",
     "`requirements-analyst` bundle. Keep this boundary clean: never open",
-    "`req:scan`, `req:draft`, `req:trace`, `req:review`, or `bcm:*`",
-    "issues from this pipeline. The only follow-ups you open are the",
+    "`req:scan`, `req:draft-trace`, `req:review`, or `bcm:*` issues",
+    "from this pipeline. The only follow-ups you open are the",
     "`req:write` issues documented below.",
     "",
     "Follow your project's shared agent conventions (`AGENTS.md`,",
@@ -12603,8 +12726,8 @@ var requirementsReviewerSubAgent = {
     "doc and must be rewritten or removed). Route follow-up issues to",
     "`req:write` so the `requirements-writer` \u2014 the only agent scoped",
     "to edit under `<REQUIREMENTS_ROOT>` \u2014 picks them up. Do **not**",
-    "route Check 4 findings to `req:trace`; the analyst bundle does",
-    "not edit requirement documents and the issue would sit",
+    "route Check 4 findings to `req:draft-trace`; the analyst bundle",
+    "does not edit requirement documents and the issue would sit",
     "unresolved. See the Follow-Up Issues section below for the full",
     "Check 3 vs. Check 4 routing split.",
     "",
@@ -13023,11 +13146,12 @@ var requirementsReviewerSubAgent = {
     "    docs, competitive-analysis notes, product-roadmap entries,",
     "    meeting extracts \u2014 that should link forward to an existing",
     "    requirement (Check 3: orphaned requirements whose fix lives",
-    "    in the source tree, not the requirement doc) \u2192 `req:trace`",
+    "    in the source tree, not the requirement doc) \u2192 `req:draft-trace`",
     "    (the `requirements-analyst` agent picks it up to backfill",
-    "    source-doc traceability). **Do not** route Check 4 broken",
-    "    cross-references here \u2014 the analyst bundle never writes to",
-    "    `<REQUIREMENTS_ROOT>` and the issue would sit unresolved.",
+    "    source-doc traceability inside its combined draft-trace phase).",
+    "    **Do not** route Check 4 broken cross-references here \u2014 the",
+    "    analyst bundle never writes to `<REQUIREMENTS_ROOT>` and the",
+    "    issue would sit unresolved.",
     "  - A new requirement is needed (e.g., a `Proposed` ADR is missing",
     "    for a deferred technology choice) \u2192 `req:scan` (the analyst",
     "    runs a scoped scan to confirm the gap, then drafts the new",
@@ -13041,7 +13165,7 @@ var requirementsReviewerSubAgent = {
     "",
     "**Check 12 exception \u2014 stale `Proposed` ADR/TR triage.** Check 12",
     "findings are the one case where follow-up flows through a new",
-    "`req:review` issue rather than `req:write` / `req:trace` /",
+    "`req:review` issue rather than `req:write` / `req:draft-trace` /",
     "`req:scan`. The finding is a decision a human must make, not an",
     "edit to a requirement document, so:",
     "",
@@ -13481,8 +13605,9 @@ var reviewRequirementsSkill = {
     "- One verification script under `<REVIEW_REPORTS_ROOT>` (only",
     "  for audits >10 documents)",
     "- Follow-up GitHub issues for every Critical and Warning finding,",
-    "  carrying the appropriate phase label (`req:write`, `req:trace`,",
-    "  or `req:scan`) for the downstream agent that should act",
+    "  carrying the appropriate phase label (`req:write`,",
+    "  `req:draft-trace`, or `req:scan`) for the downstream agent that",
+    "  should act",
     "- An aggregated `req:review` follow-up issue labeled",
     "  `status:needs-attention` when Check 12 surfaces stale `Proposed`",
     "  ADR/TR documents (one issue per review, not one per document);",
@@ -13637,9 +13762,9 @@ var requirementsReviewerBundle = {
         "The review phase produces **review reports and follow-up issues",
         "only** \u2014 it never edits requirement documents, capability",
         "models, or research notes. It files follow-up issues with the",
-        "appropriate phase label (`req:write`, `req:trace`, `req:scan`,",
-        "or \u2014 only for Check 12 stale-decision triage \u2014 `req:review`)",
-        "for the downstream agent or human to pick up.",
+        "appropriate phase label (`req:write`, `req:draft-trace`,",
+        "`req:scan`, or \u2014 only for Check 12 stale-decision triage \u2014",
+        "`req:review`) for the downstream agent or human to pick up.",
         "",
         "For audits covering more than 10 documents, the reviewer writes",
         "a read-only Python verification script alongside the report and",
@@ -13682,9 +13807,9 @@ var requirementsReviewerBundle = {
         "Authoring remains the responsibility of the `requirements-writer`",
         "(requirement docs) and `bcm-writer` (capability models) bundles.",
         "Gap discovery is the responsibility of the `requirements-analyst`",
-        "bundle. The reviewer never opens `req:scan`, `req:draft`,",
-        "`req:trace`, `req:review`, or `bcm:*` issues \u2014 only `req:write`",
-        "follow-ups from either phase.",
+        "bundle. The reviewer never opens `req:scan`, `req:draft-trace`,",
+        "`req:review`, or `bcm:*` issues \u2014 only `req:write` follow-ups",
+        "from either phase.",
         "",
         "**Soft dependency on the `requirements-writer` bundle.** The",
         "reviewer reads category templates from",
@@ -16260,18 +16385,24 @@ function renderMeetingTypes(types, agendaTemplateRoot) {
     "This project declares a meeting-type taxonomy through",
     "`AgentConfigOptions.meetings.meetingTypes`. When classifying a",
     "meeting transcript, pick the `id` from this list that best fits the",
-    "meeting; do **not** invent new type identifiers. Agenda-template",
-    `paths below are resolved relative to \`${resolvedRoot}\`.`,
-    "",
-    "| ID | Label | Scope | Cadence | Default duration | Agenda template |",
-    "|----|-------|-------|---------|------------------|-----------------|"
+    "meeting; do **not** invent new type identifiers. The `Kind` column",
+    "maps each concrete `id` onto the generic meeting-kind taxonomy",
+    "(`planning` / `review` / `brainstorm` / `standup` / `external` /",
+    "`other`) documented in the **Meeting type handling** section \u2014",
+    "apply the type-specific rules listed there based on this kind.",
+    "Agenda-template paths below are resolved relative to",
+    `\`${resolvedRoot}\`.`,
+    "",
+    "| ID | Label | Kind | Scope | Cadence | Default duration | Agenda template |",
+    "|----|-------|------|-------|---------|------------------|-----------------|"
   ];
   for (const type of types) {
+    const kind = type.kind ? `\`${type.kind}\`` : "\u2014";
     const cadence = type.cadence ? `\`${type.cadence}\`` : "\u2014";
     const duration = type.defaultDurationMinutes !== void 0 ? `${type.defaultDurationMinutes} min` : "\u2014";
     const template = type.agendaTemplatePath ? `\`${type.agendaTemplatePath}\`` : "\u2014";
     lines.push(
-      `| \`${type.id}\` | ${type.label} | \`${type.scope}\` | ${cadence} | ${duration} | ${template} |`
+      `| \`${type.id}\` | ${type.label} | ${kind} | \`${type.scope}\` | ${cadence} | ${duration} | ${template} |`
     );
   }
   return lines.join("\n");
@@ -16282,10 +16413,12 @@ function renderMeetingAreas(areas) {
     "",
     "This project declares a meeting-area routing map through",
     "`AgentConfigOptions.meetings.meetingAreas`. When a meeting note's",
-    "frontmatter carries an `area:` value matching an `id` below, route",
-    "phase-4 direct edits into the corresponding sub-tree of the docs",
-    "root (`AgentPathsConfig.docsRoot`). Meetings whose `area` does not",
-    "match any declared entry fall back to the default meetings root.",
+    "frontmatter carries an `areas:` list, route phase-4 direct edits",
+    "into the corresponding sub-trees of the docs root",
+    "(`AgentPathsConfig.docsRoot`). See the **Areas filtering** section",
+    "of the `meeting-analyst` agent for the full gating contract.",
+    "Meetings whose `areas` list is empty or does not match any",
+    "declared entry fall back to the default meetings root.",
     "",
     "| Area ID | Label | Doc-root sub-folder |",
     "|---------|-------|---------------------|"