@codyswann/lisa 2.189.6 → 2.189.8

package/plugins/lisa/skills/lisa-github-prd-intake/SKILL.md

@@ -0,0 +1,432 @@
+---
+name: lisa-github-prd-intake
+description: "Scans a GitHub repository for issues carrying the configured `ready` PRD label and runs the first eligible one through the dry-run validation pipeline. A PRD that passes every gate gets tickets written (to whatever destination tracker is configured — JIRA, GitHub Issues itself, or Linear) and the label flipped to the configured `ticketed` label; a PRD that fails gets clarifying-question comments and the label flipped to the configured `blocked` label. The GitHub counterpart of lisa-notion-prd-intake / lisa-confluence-prd-intake / lisa-linear-prd-intake. Composes existing skills (github-to-tracker, tracker-validate, tracker-source-artifacts, product-walkthrough)."
+allowed-tools: ["Skill", "Bash"]
+---
+
+# GitHub PRD Intake: $ARGUMENTS
+
+`$ARGUMENTS` is one of:
+
+- A GitHub `org/repo` token (e.g., `acme/product-prds`) — scans the repo for issues carrying the configured `ready` PRD label.
+- A full GitHub repo URL (e.g., `https://github.com/acme/product-prds`).
+- The literal token `github` — falls back to `.lisa.config.json` (`github.org` / `github.repo`).
+
+Run one intake cycle against that repo. The first eligible issue with the `ready` label is claimed, validated, routed to either the `blocked` label (with clarifying comments) or the `ticketed` label (with destination tickets created), then the cycle exits. Remaining ready PRDs stay queued for later scheduler invocations.
+
+## Workflow resolution
+
+PRD label names are read from `.lisa.config.json` `github.labels.prd.*`, falling back to defaults documented in the `config-resolution` rule. Bash pattern:
+
+```bash
+# Read role with default fallback. Local overrides global per-key.
+read_role() {
+  local role="$1" default="$2"
+  local local_v global_v
+  local_v=$(jq -r ".github.labels.prd.${role} // empty" .lisa.config.local.json 2>/dev/null)
+  global_v=$(jq -r ".github.labels.prd.${role} // empty" .lisa.config.json 2>/dev/null)
+  echo "${local_v:-${global_v:-$default}}"
+}
+
+READY=$(read_role ready "prd-ready")
+IN_REVIEW=$(read_role in_review "prd-in-review")
+BLOCKED=$(read_role blocked "prd-blocked")
+TICKETED=$(read_role ticketed "prd-ticketed")
+SHIPPED=$(read_role shipped "prd-shipped")
+```
+
+In prose below, the role names refer to the resolved labels: e.g. "the `ready` label" means whatever `github.labels.prd.ready` resolves to (default: `prd-ready`).
+
+This skill is the GitHub counterpart of `lisa-notion-prd-intake`, `lisa-confluence-prd-intake`, and `lisa-linear-prd-intake`. Phases, gates, comment templates, and rules are identical — the only differences are (1) the lifecycle is encoded as **issue labels** (mirroring Linear's project labels and Confluence's page labels), (2) the fetch / update tools are the `gh` CLI, and (3) clarifying-question comments land directly on the source PRD issue (because GitHub Issues *do* have native comments — no sentinel issue required, unlike Linear). Keep all four skills behaviorally aligned: when changing intake logic, change them together.
+
+The **PRD shipped rollup phase (3f)** transitions a `$TICKETED` PRD to `$SHIPPED` once all its generated top-level work is terminal, per the `prd-lifecycle-rollup` rule. This phase is GitHub-only here because its vendor surface (issue close + labels via `gh`) is GitHub-specific; the Linear / Confluence / Notion intake skills carry the **same** vendor-neutral rollup with their own surfaces (sibling sub-task #584, now landed). All four intake skills are behaviorally aligned across the rollup phase too — keep them in sync when changing rollup logic.
+
+## Confirmation policy
+
+Do NOT ask the caller whether to proceed. Once invoked with a repo, run the cycle to completion for the first eligible PRD — claim, validate, branch to `$BLOCKED` or `$TICKETED`, write the summary, and exit. The caller has already authorized the run by invoking the skill; re-prompting defeats the purpose of a background queue.
+
+Specifically forbidden:
+
+- Previewing projected scope and asking whether to continue.
+- Offering A/B/C-style choices like "proceed / skip / dry-run only" — the documented behavior IS the default.
+- Pausing because a PRD looks large, has many open questions, or is likely to end in `$BLOCKED`. The `blocked` label is a valid terminal state of this lifecycle, not a failure mode.
+- Pausing because the dry-run validation looks expensive.
+
+The only legitimate reasons to stop early:
+
+- Missing repo argument or required configuration. Surface and exit.
+- Repo unreachable, or the labelling convention not yet adopted (no issue carries any of `$READY` / `$IN_REVIEW` / `$BLOCKED` / `$TICKETED`). Surface and exit.
+- Empty ready set. Exit cleanly with `"No GitHub issues labeled $READY in <org>/<repo>. Nothing to do."`
+
+## Lifecycle assumed
+
+The PRD lifecycle is encoded as **issue labels**:
+
+```text
+draft → ready → in_review → blocked | ticketed → shipped → verified
+        (product)  (us)      (us)                  (product)  (product)
+```
+
+(Defaults: `prd-draft` / `prd-ready` / `prd-in-review` / `prd-blocked` / `prd-ticketed` / `prd-shipped` / `prd-verified`.)
+
+`verified` is the terminal state after `shipped`: it means the shipped product has been empirically checked against the PRD (set by `/lisa:verify-prd`, not by this intake skill). A failed post-ship verification does **not** use `blocked`; `/lisa:verify-prd` re-opens the PRD `shipped → ticketed` and creates build-ready fix tickets that auto-build and trigger a re-verify (the self-healing loop), introducing no `verifying` / `verification-failed` state. Like `draft` and `shipped`, `verified` is **product-owned** — this intake skill never sets, clears, or otherwise touches it. See the "PRD-level verification vs ticket verification" section of the `prd-lifecycle-rollup` rule.
+
+Exactly one of these labels is expected on a PRD issue at any time.
+
+This skill transitions:
+
+- `$READY` → `$IN_REVIEW` (claim)
+- `$IN_REVIEW` → `$BLOCKED` (gate failures or coverage gaps)
+- `$IN_REVIEW` → `$TICKETED` (success)
+- `$TICKETED` → `$BLOCKED` (post-write coverage gaps from Phase 3e)
+- `$TICKETED` → `$SHIPPED` (PRD shipped rollup, Phase 3f — only when **all** generated top-level children are terminal)
+
+The `draft` and `verified` labels are owned by product and are never touched here (`verified` is set by `/lisa:verify-prd` after empirical PRD-level acceptance). The `shipped` label is set by this skill's **rollup phase (3f)** when, and only when, the PRD's generated top-level work is all terminal — per the `prd-lifecycle-rollup` rule; product may also set it by hand. Rollup never advances a PRD to `shipped` on partial completion, and never closes a PRD issue at shipped. `/lisa:verify-prd` closes the issue only after a verified PASS.
+
+A "transition" means: remove the old lifecycle label and add the new one (`gh issue edit <num> --remove-label <old> --add-label <new>`). The skill MUST verify exactly one lifecycle label is present after the update.
+
+If the repo has not yet adopted these labels, this skill cannot run. See "Adoption" at the bottom.
+
+**Label namespace separation:** the PRD lifecycle uses the configured PRD labels (defaults `prd-*`). The build-queue lifecycle (used by `lisa-github-build-intake`) uses the configured build labels (defaults `status:*`). The two never overlap. When the destination tracker is also GitHub Issues (self-host case), the same repo can host both — but a single issue is either a PRD (carrying a PRD lifecycle label) or a build ticket (carrying a build label), never both.
+
+## Phases
+
+### Phase 1 — Resolve the repo
+
+1. Parse `$ARGUMENTS`:
+   - `org/repo` token → use as-is.
+   - GitHub URL → extract `org` and `repo`.
+   - Literal `github` → resolve from `.lisa.config.json`; error if not set.
+2. Confirm `gh auth status` succeeds.
+3. Confirm the repo is reachable: `gh repo view <org>/<repo> --json name`.
+4. Verify the PRD label set exists:
+   ```bash
+   gh label list --repo <org>/<repo> --json name --jq '.[] | .name' \
+     | grep -xE "$READY|$IN_REVIEW|$BLOCKED|$TICKETED|$SHIPPED"
+   ```
+   If none of the configured PRD labels are present, surface a label-convention error and exit (see "Adoption").
+
+### Phase 2 — Find ready PRDs
+
+```bash
+gh issue list --repo <org>/<repo> --label "$READY" --state open --limit 100 \
+  --json number,title,body,labels,author,milestone,createdAt,updatedAt,url
+```
+
+For each candidate, confirm exactly one lifecycle label is present (the `--label` filter selects `$READY` matches, but a PRD could have ended up with two labels by hand — that's a misconfiguration, not a normal queue entry).
+
+If empty, run a secondary check:
+
+```bash
+gh issue list --repo <org>/<repo> --state open --limit 100 --json number,labels \
+  --jq "[.[] | .labels[] | select(.name == \"$READY\" or .name == \"$IN_REVIEW\" or .name == \"$BLOCKED\" or .name == \"$TICKETED\") | .name] | unique"
+```
+
+If no PRD lifecycle labels appear on any open issue → convention not adopted; surface error and exit. If lifecycle labels exist but none are `$READY` → genuinely empty queue, exit cleanly with the idle message.
+
+### Phase 3 — Process the first eligible ready PRD
+
+Select the first ready PRD issue returned by Phase 2 and process only that issue. Later scheduler invocations process the remaining ready PRDs.
+
+#### 3a. Claim
+
+```bash
+gh issue edit <num> --repo <org>/<repo> --remove-label "$READY" --add-label "$IN_REVIEW"
+```
+
+This is the idempotency lock — a re-entrant cycle's `--label $READY` filter won't see this issue again.
+
+If the relabel fails (permission, race), log and skip. Do not proceed to validation on a PRD you didn't successfully claim.
+
+This skill never edits the PRD body. Communication with product happens only through comments.
+
+#### 3b. Dry-run validation
+
+Invoke the `lisa-github-to-tracker` skill with `dry_run: true` and the PRD issue ref. The skill returns a structured report containing:
+- The planned ticket hierarchy
+- Per-ticket validation verdicts and remediation
+- An overall PASS / FAIL verdict
+- A failure count
+
+This call indirectly invokes `lisa-tracker-source-artifacts` (artifact extraction + classification) and `lisa-product-walkthrough` (when the PRD touches existing user-facing surfaces). All gate logic lives in `lisa-tracker-validate` (which dispatches to `lisa-jira-validate-ticket` or `lisa-github-validate-issue` depending on the configured destination).
+
+#### 3c. Branch on the verdict
+
+**If `PASS`** (every planned ticket passed every applicable gate):
+
+1. Re-invoke `lisa-github-to-tracker` with `dry_run: false` to actually write tickets. This re-runs the planning phases and the preservation gate.
+2. Capture the created ticket refs.
+3. Post a comment on the PRD issue listing the created tickets with their URLs:
+   ```bash
+   gh issue comment <prd-num> --repo <org>/<repo> --body-file /tmp/ticketed-comment.md
+   ```
+   Lead with: `"Ticketed by Claude. Created N tickets in <destination> — see below. Add the $SHIPPED label after the work is delivered."` The destination is named (JIRA / GitHub Issues) so product knows where to look.
+4. Transition labels: `gh issue edit <prd-num> --remove-label "$IN_REVIEW" --add-label "$TICKETED"`.
+5. **Run Phase 3e (coverage audit)** before considering this PRD done.
+
+**If `FAIL`**:
+
+The audience is the **product team**, not engineers. Follow the strict comment rules below.
+
+##### 3c.1 Partition failures
+
+1. Drop every failure where `product_relevant = false`. Internal data-quality problems — the agent should fix its own spec, not ask product. Record dropped failures under `Errors` in the cycle summary.
+2. Group remaining product-relevant failures by `prd_anchor` (a section heading or a `selection_with_ellipsis` snippet from the PRD body). Failures sharing an anchor become one comment.
+
+##### 3c.2 Render each comment
+
+GitHub Issues do not have selection-anchored comments (unlike Confluence inline comments). Approximate by quoting the relevant body excerpt at the top of each comment:
+
+```bash
+gh issue comment <prd-num> --repo <org>/<repo> --body-file /tmp/anchored-comment-N.md
+```
+
+Each comment template MUST contain these parts in this order, no exceptions:
+
+```text
+[<Category badge>] <prd_section heading text>
+
+> <quoted excerpt from the PRD body, ~10–30 words around the anchor — this stands in for inline anchoring>
+
+**What's unclear:** <validator's `what` field, verbatim — already product-readable>
+
+**Recommendation:** <validator's `recommendation` field, verbatim — must contain 1–3 concrete options, never a generic "please clarify">
+
+**Action:** Update this section in the PRD, then replace the `$BLOCKED` label with `$READY` on the issue and Claude will re-run intake.
+```
+
+If multiple failures share an anchor, render each as its own `**What's unclear:** ... **Recommendation:** ...` block within the same comment, separated by `---`. Keep the single `[Category badge]` heading at the top.
+
+For unanchored failures (`prd_anchor: null`), post one rollup comment prefixed with `Issues without a specific section anchor:`.
+
+##### 3c.3 Category badges
+
+| Validator category | Badge label |
+|---------------------|-------------|
+| `product-clarity` | `[Product clarity]` |
+| `acceptance-criteria` | `[Acceptance criteria]` |
+| `design-ux` | `[Design / UX]` |
+| `scope` | `[Scope]` |
+| `dependency` | `[Dependency]` |
+| `data` | `[Data]` |
+| `technical` | `[Technical]` |
+
+`structural` failures must never reach this step (filtered in 3c.1).
+
+##### 3c.4 Forbidden in product comments
+
+- Gate IDs (`S4`, `F2`, etc.).
+- GitHub-Issues-specific terminology (`gh`, `sub-issue`, `label namespace`) — paraphrase before posting.
+- Internal skill names (`lisa-tracker-validate`, `github-to-tracker`).
+- Engineering shorthand (`AC`, `OOS`, `repo`, `env var`).
+- "Clarify this" / "Please specify" without candidate resolutions.
+
+##### 3c.5 Label transition
+
+After all comments are posted, transition: `gh issue edit <num> --remove-label "$IN_REVIEW" --add-label "$BLOCKED"`. Do NOT write any tickets.
+
+#### 3d. Stop
+
+Stop immediately after the claimed PRD is ticketed, blocked, or recorded as an error.
+
+#### 3e. Coverage audit (mandatory after $TICKETED)
+
+Per-ticket gates prove each ticket is well-formed; they do NOT prove the *set* of tickets covers the *whole* PRD. Invoke `lisa-prd-ticket-coverage` to catch silent drops.
+
+1. Invoke `lisa-prd-ticket-coverage` with `<PRD URL> tickets=[<created refs from 3c step 2>]`. The coverage skill auto-detects the PRD vendor from the URL host (`github.com` → GitHub).
+2. Read the verdict:
+
+   | Verdict | Action |
+   |---------|--------|
+   | `COMPLETE` | Done. Leave label as `$TICKETED`. End the cycle. |
+   | `COMPLETE_WITH_SCOPE_CREEP` | Post an advisory comment on the PRD issue naming the scope-creep tickets. Leave label as `$TICKETED`. |
+   | `GAPS_FOUND` | The created ticket set is incomplete. (a) For each gap, post a comment using the same product-facing template as Phase 3c.2 — anchored when `prd_anchor` is non-null. (b) Post one summary comment listing the tickets that *were* successfully created. (c) Transition labels from `$TICKETED` back to `$BLOCKED`. |
+   | `NO_TICKETS_FOUND` | Should not happen if step 2 succeeded. Log as Error; leave label as `$TICKETED` with a flag comment. |
+
+3. The created tickets remain in the destination tracker regardless of the verdict. The audit only tells us whether *more* are needed.
+
+#### 3f. PRD shipped rollup
+
+A PRD's lifecycle terminal state (`shipped`) is **derived** from whether the work it generated is done — it is never set by hand here on its own authority. This phase implements the GitHub leg of that derivation, per the `prd-lifecycle-rollup` rule (cite it by slug; do not restate its taxonomy or terminal-state semantics here). Linear / Confluence / Notion rollup is a sibling sub-task (#584) and is out of scope for this skill.
+
+Rollup runs over PRD issues that are already `$TICKETED` (the only state from which a PRD can ship): the freshly-ticketed PRD from Phase 3c, and — because rollup also catches PRDs whose children finished in a *later* cycle — every issue currently carrying `$TICKETED`. Process each independently; one PRD never blocks another's rollup.
+
+##### 3f.0 Shipped remains open for verification
+
+There is no close/archive configuration at the shipped hop. Rollup sets `$SHIPPED` and leaves the PRD issue **open** so Phase 3g can dispatch `/lisa:verify-prd`. Provider-native issue closure is owned by `/lisa:verify-prd` after it transitions `$SHIPPED → verified` on a PASS.
+
+##### 3f.1 Idempotency guard (no-op if already shipped)
+
+Rollup is keyed by the PRD's current state. If the PRD already carries `$SHIPPED`, it is a **no-op** — do not re-transition, do not close, do not re-comment. Record it as `already shipped (no-op)` in the cycle summary and move on. This is what makes re-running intake safe.
+
+##### 3f.2 Read the generated top-level child set
+
+Read the PRD's **generated top-level work** — its created Epics and any top-level Stories created directly under it, **excluding** leaf Sub-tasks and any Story nested under a generated Epic (`prd-lifecycle-rollup` rule, generated-top-level-work contract). Use two sources, native first:
+
+1. **Native sub-issues (primary).** Traverse the PRD issue's native sub-issue graph via the GraphQL `subIssues` query (the same query `lisa-github-read-issue` Phase 3 uses). The PRD's direct `subIssues` nodes are its top-level children:
+
+   ```bash
+   gh api graphql -f query='
+   query($org:String!,$repo:String!,$number:Int!){
+     repository(owner:$org,name:$repo){
+       issue(number:$number){
+         subIssues(first: 100) {
+           nodes {
+             number title state url
+             repository { nameWithOwner }
+             labels(first: 50) { nodes { name } }
+           }
+         }
+       }
+     }
+   }' -F org=<org> -F repo=<repo> -F number=<prd-num>
+   ```
+
+2. **Documented `## Tickets` section (fallback).** When native sub-issues are unavailable (older GHES, sub-issues feature off, or the source PRD and the destination tracker are different systems so the children were never linked as sub-issues), parse the machine-readable generated-work section `lisa-prd-backlink` writes to the PRD body (`## Tickets`, alias `## Generated Work`; see #582). Top-level children are the `### <Epic key>: <title>` group headers' first line (`- [<ref>](<url>) — Epic`) plus any top-level Story listed directly under `### Unparented items`. Lines nested deeper (`  - ... — Story:` under an Epic, `    - ... — Sub-task:`) are descendants, NOT top-level children — skip them.
+
+   ```bash
+   # Top-level child refs = Epic lines (top indent) + Unparented top-level Stories.
+   # Sub-tasks and Stories nested under an Epic are descendants — excluded.
+   gh issue view <prd-num> --repo <org>/<repo> --json body --jq '.body' \
+     | awk '/^## (Tickets|Generated Work)/{insec=1;next} /^## /{insec=0}
+            insec && /^- \[.*\] — Epic/{print}
+            insec && /^### Unparented items/{unp=1;next}
+            insec && unp && /^- \[.*\] — Story/{print}'
+   ```
+
+Dedupe the resulting child set by **child-ref identity** (`owner/repo#number`) so a child that appears both as a native sub-issue and in the documented section is counted once (`prd-lifecycle-rollup` idempotency dedupe key). If neither source yields any child (the PRD generated nothing, or the relationship was never recorded), record `no generated top-level children — rollup skipped` and leave the PRD as `$TICKETED`; do not ship an empty PRD.
+
+##### 3f.3 Apply the terminal-state predicate
+
+For each top-level child, fetch its state + labels (already present from the GraphQL nodes, or `gh issue view <child-num> --json state,labels`) and classify per the `prd-lifecycle-rollup` GitHub predicate:
+
+- **Terminal (shipped).** The child issue is **CLOSED** *and* (where the build-status label is in use) carries the resolved build `done` role label (`status:done` by default). A child Epic is terminal only when it has itself rolled up to its own terminal state per `leaf-only-lifecycle` — read the child's own resolved state; do not re-derive it from its leaves here.
+- **Terminal-but-dropped.** The child is closed **as not planned** (`stateReason == "not_planned"`). It does not hold the PRD open and is excluded from the shipped set — treated like a won't-do leaf.
+- **Incomplete / blocked.** Anything else: still open, or closed without the `done` label. Holds the PRD open.
+
+The set of **required** children for the all-terminal check is the top-level children minus the terminal-but-dropped ones.
+
+##### 3f.4 Branch on the rollup verdict
+
+**All required children terminal** (every required top-level child is terminal; at least one required child exists):
+
+1. Transition labels: `gh issue edit <prd-num> --repo <org>/<repo> --remove-label "$TICKETED" --add-label "$SHIPPED"`. Verify exactly one lifecycle label remains (the single-label invariant).
+2. Leave the PRD issue open for `/lisa:verify-prd`; do not close at the shipped hop.
+3. Post a short rollup comment naming the terminal child set and (when dropped children exist) the dropped set, so the audit trail records *why* the PRD shipped. Lead with `"Shipped by Claude — all generated top-level work is complete."`
+
+**Any required child incomplete / blocked**:
+
+1. Leave the PRD label as `$TICKETED` and leave the issue **open**. Do NOT add `$SHIPPED`. Do NOT close.
+2. Report the incomplete child set — both in the cycle summary and, when at least one cycle has previously ticketed this PRD, as a single advisory comment listing the still-open children (`- <ref> "<title>" — <state>`), so product can see what's blocking the rollup. Keep it idempotent: regenerate the advisory rather than appending a fresh one each cycle.
+
+##### 3f.5 Rollup is GitHub-only and cites the rule
+
+This phase only touches GitHub PRD issues. It implements exactly one PRD-lifecycle hop — `$TICKETED → $SHIPPED` — and deliberately leaves native closure to `/lisa:verify-prd` after `$SHIPPED → verified`. All terminal-state semantics, the generated-top-level-work boundary, the env-keyed `done` resolution, and the dedupe-by-child-ref idempotency come from the `prd-lifecycle-rollup` rule; this skill is its GitHub implementation, not a second source of truth.
+
+#### 3g. PRD verification dispatch (close the loop on shipped PRDs)
+
+`shipped` and `verified` are distinct facts about a PRD (see the `prd-lifecycle-rollup` rule's "PRD-level verification vs ticket verification" and "Closing the loop" sections). Rollup (3f) only reaches `$SHIPPED`; the `shipped → verified` (pass) / `shipped → ticketed` (fail) hops are owned by `/lisa:verify-prd`. This phase **closes that loop** by dispatching the initiative-level acceptance gate for shipped PRDs. It never performs the verification transition itself — the "never sets the verification outcome" invariant holds: `lisa-verify-prd`, not this skill, sets `verified` (or, on failure, re-opens the PRD to `ticketed`).
+
+Re-query the PRDs currently carrying `$SHIPPED` via `gh issue list --repo <org>/<repo> --label "$SHIPPED" --state open`. Pick the **first** one and invoke `lisa-verify-prd <issue-url>`. Process **one shipped PRD per cycle** — `lisa-verify-prd` is a heavy full flow (spec-conformance + empirical verification + fix-issue creation), so it is bounded exactly like the single-ready-PRD claim in Phase 3; the scheduler drains the rest.
+
+**Per-cycle combined bound:** each scheduler cycle dispatches at most one ready PRD (the Phase 3 single-ready-PRD claim) **and** at most one shipped PRD for verification (this Phase 3g dispatch), for a maximum of two PRD operations per cycle. Ready intake runs first (Phase 3), then shipped verify (Phase 3g).
+
+`lisa-verify-prd` owns the outcome: on a CONFORMS verdict with all empirical checks passing it transitions `$SHIPPED → verified` and posts evidence; on a conformance miss or a failing/unavailable check it **re-opens the PRD `$SHIPPED → ticketed`** (never `blocked`) and creates **build-ready** fix tickets registered as the PRD's generated work, then posts a failure report — the fix tickets auto-build, rollup (3f) re-ships the PRD once they are terminal, and a later cycle re-verifies (the self-healing loop). Either branch moves the PRD out of `$SHIPPED`, so it is not re-picked this cycle; a PRD whose generated work is not actually terminal is guard-stopped by `lisa-verify-prd` (left `$SHIPPED`) — that is verify-prd's gate, not this skill's.
+
+This phase, like 3f, is **behaviorally identical across all four intake skills** (`github-prd-intake`, `linear-prd-intake`, `notion-prd-intake`, `confluence-prd-intake`) — only the `$SHIPPED` query surface differs; keep them aligned. Record the dispatched PRD + verify-prd's verdict in the summary.
+
+### Phase 4 — Summary report
+
+```text
+## github-prd-intake summary
+
+Repo: <org>/<repo> (<URL>)
+Cycle started: <ISO timestamp>
+Cycle completed: <ISO timestamp>
+
+PRDs processed: <n>
+- $TICKETED: <n>
+  - <issue-ref> "<title>" → <epic-ref> + <story-count> stories + <subtask-count> sub-tasks (coverage: COMPLETE | COMPLETE_WITH_SCOPE_CREEP)
+- $BLOCKED: <n>
+  - <issue-ref> "<title>" → <gate-failure-count> gate failures (pre-write) OR <gap-count> coverage gaps (post-write)
+- Errors (claim failed, etc): <n>
+  - <issue-ref> "<title>" — <reason>
+
+Rollup (Phase 3f):
+- $SHIPPED: <n>
+  - <issue-ref> "<title>" → all <child-count> top-level children terminal (<dropped-count> dropped); left open for verify-prd
+- Held open (incomplete children): <n>
+  - <issue-ref> "<title>" → <incomplete-count> of <child-count> top-level children still open
+- Already shipped (no-op): <n>
+- No generated children (rollup skipped): <n>
+
+Total tickets created: <n>
+Coverage audit summary: <n> COMPLETE / <n> COMPLETE_WITH_SCOPE_CREEP / <n> GAPS_FOUND
+```
+
+Print to the agent's output. Do not write this summary to GitHub.
+
+## Self-host edge case (PRD repo = destination repo)
+
+When the configured destination tracker is GitHub Issues AND the PRD repo is the same as the destination repo, both reads and writes hit the same place. Disambiguation rules:
+
+- A PRD issue carries a PRD lifecycle label (configured under `github.labels.prd.*`). Built tickets carry a `type:*` + build-status label set (configured under `github.labels.build.*`), but never a PRD lifecycle label.
+- The "Ticketed by Claude" comment on the PRD links to the destination ticket numbers (which live in the same repo, so the links are simple `#<n>` refs).
+- `lisa-prd-ticket-coverage` filters out the source PRD itself when listing destination tickets — the PRD is never a ticket of its own work.
+
+## Idempotency & safety
+
+- **One item per cycle**: this skill processes the first eligible ready PRD issue from Phase 2, then exits. New or remaining ready issues are picked up by later scheduler invocations.
+- **No writes outside the lifecycle**: this skill only ever writes to the destination tracker via `lisa-github-to-tracker` (which delegates to `lisa-tracker-write`), only ever changes labels among `$IN_REVIEW`, `$BLOCKED`, `$TICKETED`, `$SHIPPED`, only ever comments on the source PRD issue. It never edits PRD bodies, never touches the `draft` label, never closes PRD issues at the shipped hop, and never deletes any issue.
+- **Claim-first ordering**: the label flip to `$IN_REVIEW` happens BEFORE validation runs.
+- **Failure handling**: an exception processing the selected PRD is caught and recorded under "Errors" in the summary, then the cycle exits. The PRD that errored is left labeled `$IN_REVIEW` — humans investigate from there.
+- **Single-label invariant**: after every transition, verify exactly one lifecycle label is present.
+- **Rollup idempotency**: rollup (Phase 3f) is a no-op on a PRD already carrying `$SHIPPED` — no duplicate transition, no shipped-time close, no duplicate comment. The all-terminal condition is a pure function of the children's current states, so recomputing it is safe to re-run. Native closure only follows verified PASS in `/lisa:verify-prd`.
+
+## Configuration
+
+Same configuration as `lisa-github-to-tracker`. See that skill for the full table. Key items:
+
+- **From `.lisa.config.json`**: `github.org` and `github.repo` (required for the source repo, and also for the destination repo when `tracker = "github"` — self-host case), plus `github.labels.prd.*` for the lifecycle label vocabulary (all optional; defaults documented above).
+- **From environment variables**: `E2E_BASE_URL`, `E2E_TEST_PHONE`, `E2E_TEST_OTP`, `E2E_TEST_ORG`, `E2E_GRAPHQL_URL` (operational E2E test config).
+
+Destination tracker config (jira / github / linear) is consumed by `lisa-tracker-write` internally — this skill does NOT read it. If any required value is missing, surface and exit — never invent values.
+
+| Field | Default | Purpose |
+|-------|---------|---------|
+| `.lisa.config.json` `github.labels.prd.ready` | `prd-ready` | Label signalling "PRD ready for ticketing" |
+| `.lisa.config.json` `github.labels.prd.in_review` | `prd-in-review` | Label set on claim |
+| `.lisa.config.json` `github.labels.prd.blocked` | `prd-blocked` | Label set on validation failure |
+| `.lisa.config.json` `github.labels.prd.ticketed` | `prd-ticketed` | Label set on success |
+| `.lisa.config.json` `github.labels.prd.shipped` | `prd-shipped` | Label product sets after delivery |
+
+## Rules
+
+- Never write to the destination tracker outside of `lisa-github-to-tracker` → `lisa-tracker-write`.
+- Never add or remove a label this skill doesn't own (`$IN_REVIEW`, `$BLOCKED`, `$TICKETED`, and `$SHIPPED` via the rollup phase only). Product owns the `draft` and `ready` PRD labels; product and the rollup phase (3f) both set `shipped`.
+- Set `$SHIPPED` only from the rollup phase, and only when all generated top-level children are terminal per the `prd-lifecycle-rollup` rule. Never ship on partial completion and never close at shipped.
+- Never edit a PRD's body. Communication with product happens only via comments.
+- Never post a single dump of all gate failures on one comment. One comment per `prd_anchor` group, plus one rollup for unanchored failures.
+- Never include a gate ID, internal skill name, or engineering shorthand in a comment body.
+- Never run more than one intake cycle concurrently against the same repo.
+- If `lisa-github-to-tracker` returns errors, treat them as gate failures: comment + `$BLOCKED`. Don't silently fail.
+
+## Adoption (one-time per repo)
+
+Before this skill can run, the repo must adopt the PRD lifecycle issue-label convention. Using the defaults:
+
+1. Create the labels:
+   ```bash
+   gh label create prd-draft --color C5DEF5 --description "PRD in progress (product owns)" --repo <org>/<repo>
+   gh label create prd-ready --color FBCA04 --description "PRD ready for ticketing" --repo <org>/<repo>
+   gh label create prd-in-review --color 5319E7 --description "Claude is reviewing this PRD" --repo <org>/<repo>
+   gh label create prd-blocked --color D93F0B --description "PRD blocked — see comments" --repo <org>/<repo>
+   gh label create prd-ticketed --color 0E8A16 --description "Tickets created — see comments" --repo <org>/<repo>
+   gh label create prd-shipped --color 1D76DB --description "Work delivered (product owns)" --repo <org>/<repo>
+   ```
+   If your project overrides any `github.labels.prd.*` role name in config, substitute the actual label names you configured.
+2. Apply the `$READY` label to issues that are ready for ticketing.
+3. Reserve `$IN_REVIEW`, `$BLOCKED`, `$TICKETED` for this skill — humans should not set them manually except to recover from an error.
+4. Keep the PRD label namespace strictly separate from the build-queue label namespace owned by `lisa-github-build-intake`.
+
+If the repo hasn't adopted these labels, the first run exits with a label-convention error (not the idle empty-set message) — this distinguishes setup from a genuinely empty queue.

package/plugins/lisa/skills/lisa-github-prd-intake/agents/openai.yaml

@@ -0,0 +1,4 @@
+display_name: "Github PRD Intake"
+short_description: "Scans a GitHub repository for issues carrying the configured `ready` PRD label and runs the first eligible one through the dry-run…"
+default_prompt:
+  - "Use $lisa-github-prd-intake: Scans a GitHub repository for issues carrying the configured `ready` PRD label and runs the first eligible one through the dry-run…."