@edgehero/pi-dispatch-admin 0.5.0 → 0.6.0

package/README.md

@@ -46,7 +46,7 @@ Every trigger is one `{ on, run }` entry in a single `triggers.json`, read live
 
 Four forges: GitHub, GitLab, Forgejo (and Gitea), Azure DevOps. **Who may fire a trigger is your forge's decision, not this service's**: on GitHub the label *is* the approval, because only collaborators can apply one, while GitLab, Forgejo and Azure resolve the actor's permission through their APIs. Each forge's action vocabulary is validated when the file loads, so a word from the wrong forge is refused instead of silently never matching.
 
-**Flows, and workflows.** `run.flow` names a skill committed to the target repo at `.pi/skills/<flow>/SKILL.md`, read from the **default branch**, so the repo owns the prompt and merging it is the repo's consent. A skill may call other skills, which is already a workflow. For typed multi-stage ones, a pi extension such as `@juicesharp/rpiv-workflow` can be staged into the deployment: pinned to an exact version, installed on your host (never at job time, since jobs run offline), loaded in every container, and declinable per trigger.
+**Flows, and workflows.** `run.flow` names a skill committed to the target repo at `.pi/skills/<flow>/SKILL.md`, read from the **default branch**, so the repo owns the prompt and merging it is the repo's consent. A skill may call other skills, which is already a workflow. For typed multi-stage ones, a pi extension such as `@juicesharp/rpiv-workflow` can be staged into the deployment: pinned to an exact version, installed on your host (never at job time, since jobs run offline), loaded in every container, and declinable per trigger. Anything you already installed with `pi install` is staged automatically, at the version your host has, so there is nothing to declare twice.
 
 **How a workflow gets triggered, in one line:** `label / comment / PR / cron` fires **one job in one container**, that job runs `run.flow`, and the flow drives whatever stages follow. Four consequences worth knowing before you build on it. A trigger names a **flow, never a workflow**, so which stages run is decided inside the job by that skill. A job is **not an interactive session** (the container hands pi one assembled prompt and reads the exit line), so a workflow extension's slash command has nobody to type it: in a job the flow's instructions drive the workflow, or a small extension you also staged calls its API from a lifecycle hook. One trigger is **one job, one budget slot and one turn budget**, so ten stages share the same `PI_MAX_TURNS` and per-job token budget, and exhausting either ends the job as a policy refusal that is never retried. And whether the workflow's own state survives depends on the trigger kind: a cron or CLI job has your folder mounted read-write so state persists between runs, while a forge job gets a fresh clone that is discarded with the container. Full reference: [`docs/workflows.md`](https://github.com/edgehero/pi-dispatch/blob/main/docs/workflows.md).
 

package/dist/index.mjs

@@ -7935,11 +7935,13 @@ function normalizeCron(on, run, index, path, state) {
   }
   const packages = validatePackagesFlag(run, `cron trigger "${id}"`, path);
   const image = validateImageRef(run, `cron trigger "${id}"`, path);
+  const skillsDir = validateSkillsDir(run, `cron trigger "${id}"`, path);
+  validateInstructions(run, `cron trigger "${id}"`, path, { cron: true });
   const resume = validateResumeFlag(run, `cron trigger "${id}"`, path);
   validateReplicas(run, `cron trigger "${id}"`, path);
   return {
     on: { type: "cron", id, pattern },
-    run: { kind: "local", folder: run.folder, flow: run.flow, task: run.task, provider: run.provider, model: run.model, maxTurns: run.maxTurns, github: run.github, packages, image, resume }
+    run: { kind: "local", folder: run.folder, flow: run.flow, task: run.task, provider: run.provider, model: run.model, maxTurns: run.maxTurns, github: run.github, packages, image, resume, ...skillsDir !== void 0 && { skillsDir } }
   };
 }
 function validatePackagesFlag(run, at, path) {
@@ -7973,6 +7975,31 @@ function validateImageRef(run, at, path) {
   }
   return image;
 }
+function validateSkillsDir(run, at, path) {
+  const dir = run.skillsDir;
+  if (dir === void 0) return void 0;
+  if (typeof dir !== "string" || dir.trim() === "") {
+    throw configError(`${at}: run.skillsDir must be a non-empty string when present: ${path}`);
+  }
+  if (dir !== dir.trim()) {
+    throw configError(`${at}: run.skillsDir must not have leading or trailing whitespace (got ${JSON.stringify(dir)}): ${path}`);
+  }
+  return dir;
+}
+function validateInstructions(run, at, path, { cron = false } = {}) {
+  const text = run.instructions;
+  if (text === void 0) return void 0;
+  if (cron) {
+    throw configError(`${at}: run.instructions is not accepted on a cron trigger -- a local job's prompt IS run.task, the same operator-authored text in the same place. Put the standing instruction at the top of run.task: ${path}`);
+  }
+  if (typeof text !== "string" || text.trim() === "") {
+    throw configError(`${at}: run.instructions must be a non-empty string when present: ${path}`);
+  }
+  if (text.length > INSTRUCTIONS_MAX) {
+    throw configError(`${at}: run.instructions is ${text.length} characters, over the ${INSTRUCTIONS_MAX} cap -- a standing instruction is a sentence or two. Anything longer belongs in the flow's own SKILL.md, or in the global overlay's APPEND_SYSTEM.md if it applies to every job: ${path}`);
+  }
+  return text;
+}
 function validatePredicate(on, index, path, requirePositive) {
   const at = `trigger at index ${index}`;
   for (const key of ["any", "all", "none"]) {
@@ -8027,12 +8054,14 @@ function normalizeLabel(on, run, index, path) {
   }
   const packages = validatePackagesFlag(run, at, path);
   const image = validateImageRef(run, at, path);
+  const skillsDir = validateSkillsDir(run, at, path);
+  const instructions = validateInstructions(run, at, path);
   const resume = validateResumeFlag(run, at, path);
   const repository = validateRepository(run, "label", at, path);
   const replicas = validateReplicas(run, at, path);
   return {
     on: { type: "label", any: predicate.any, all: predicate.all, none: predicate.none },
-    run: { kind: run.kind, flow: run.flow, packages, image, resume, replicas, ...repository !== void 0 && { repository } }
+    run: { kind: run.kind, flow: run.flow, packages, image, resume, replicas, ...skillsDir !== void 0 && { skillsDir }, ...instructions !== void 0 && { instructions }, ...repository !== void 0 && { repository } }
   };
 }
 function normalizeComment(on, run, index, path, state) {
@@ -8049,12 +8078,14 @@ function normalizeComment(on, run, index, path, state) {
   }
   const packages = validatePackagesFlag(run, at, path);
   const image = validateImageRef(run, at, path);
+  const skillsDir = validateSkillsDir(run, at, path);
+  const instructions = validateInstructions(run, at, path);
   const resume = validateResumeFlag(run, at, path);
   const repository = validateRepository(run, "comment", at, path);
   const replicas = validateReplicas(run, at, path);
   return {
     on: { type: "comment", phrase: on.phrase },
-    run: { kind: run.kind, flow: run.flow, packages, image, resume, replicas, ...repository !== void 0 && { repository } }
+    run: { kind: run.kind, flow: run.flow, packages, image, resume, replicas, ...skillsDir !== void 0 && { skillsDir }, ...instructions !== void 0 && { instructions }, ...repository !== void 0 && { repository } }
   };
 }
 function normalizePullRequest(on, run, index, path) {
@@ -8070,6 +8101,7 @@ function normalizePullRequest(on, run, index, path) {
       throw configError(`${at}: pull_request on.action has an unsupported ${run.kind} action ${JSON.stringify(a)} (expected ${expected}): ${path}`);
     }
   }
+  const reviewState = validateReviewState(on, actions, run, at, path);
   const labelAction = forgeSpec(run.kind)?.prLabelAction;
   const requirePositive = typeof labelAction === "string" && actions.includes(labelAction);
   if (run.kind === "azure" && (on.any !== void 0 || on.all !== void 0 || on.none !== void 0)) {
@@ -8081,22 +8113,55 @@ function normalizePullRequest(on, run, index, path) {
   }
   const packages = validatePackagesFlag(run, at, path);
   const image = validateImageRef(run, at, path);
+  const skillsDir = validateSkillsDir(run, at, path);
+  const instructions = validateInstructions(run, at, path);
   const resume = validateResumeFlag(run, at, path);
   validateRepository(run, "pull_request", at, path);
   const replicas = validateReplicas(run, at, path);
   return {
-    on: { type: "pull_request", action: [...actions], any: predicate.any, all: predicate.all, none: predicate.none },
-    run: { kind: run.kind, flow: run.flow, packages, image, resume, replicas }
+    on: {
+      type: "pull_request",
+      action: [...actions],
+      // Absent rather than present-and-undefined: an unnarrowed rule's normalized shape must stay
+      // byte-identical to the one every pre-#66 trigger file produces.
+      ...reviewState !== void 0 && { reviewState },
+      any: predicate.any,
+      all: predicate.all,
+      none: predicate.none
+    },
+    run: { kind: run.kind, flow: run.flow, packages, image, resume, replicas, ...skillsDir !== void 0 && { skillsDir }, ...instructions !== void 0 && { instructions } }
   };
 }
-var ON_TYPES, PR_ACTIONS, ID_CHARSET, REPLICAS_MAX;
+function validateReviewState(on, actions, run, at, path) {
+  if (on.reviewState === void 0) return void 0;
+  if (run.kind !== "github") {
+    throw configError(`${at}: on.reviewState is github-only (got ${run.kind}), no other forge reports a review verdict: ${path}`);
+  }
+  if (!actions.includes(REVIEW_ACTION)) {
+    throw configError(`${at}: on.reviewState requires on.action to include ${JSON.stringify(REVIEW_ACTION)}, otherwise it narrows nothing: ${path}`);
+  }
+  if (!Array.isArray(on.reviewState) || on.reviewState.length === 0) {
+    throw configError(`${at}: on.reviewState must be a non-empty array: ${path}`);
+  }
+  const expected = [...REVIEW_STATES].join("|");
+  for (const s of on.reviewState) {
+    if (!REVIEW_STATES.has(s)) {
+      throw configError(`${at}: on.reviewState has an unsupported review state ${JSON.stringify(s)} (expected ${expected}): ${path}`);
+    }
+  }
+  return [...on.reviewState];
+}
+var ON_TYPES, PR_ACTIONS, REVIEW_STATES, REVIEW_ACTION, ID_CHARSET, REPLICAS_MAX, INSTRUCTIONS_MAX;
 var init_triggers = __esm({
   "../worker/src/triggers.mjs"() {
     init_config2();
     init_forges();
     ON_TYPES = /* @__PURE__ */ new Set(["cron", "label", "comment", "pull_request"]);
     PR_ACTIONS = {
-      github: /* @__PURE__ */ new Set(["labeled", "opened", "synchronize", "reopened"]),
+      github: /* @__PURE__ */ new Set(["labeled", "opened", "synchronize", "reopened", "review_submitted"]),
+      // GitLab's `approved` is its review gate (a member approved the MR). It is NOT github's
+      // `review_submitted` renamed: `approved` is one verdict, `review_submitted` is every verdict, which is
+      // what `on.reviewState` below exists to narrow.
       gitlab: /* @__PURE__ */ new Set(["open", "update", "reopen", "approved"]),
       // Forgejo's own spellings. `label_updated` is its `labeled` and `synchronized` its `synchronize` -- a
       // one-letter difference that an operator would otherwise discover as a trigger that loads clean and
@@ -8109,8 +8174,11 @@ var init_triggers = __esm({
       // requests -- which is why azure's `prLabelAction` is null and a predicated PR rule is refused below.
       azure: /* @__PURE__ */ new Set(["created", "updated"])
     };
+    REVIEW_STATES = /* @__PURE__ */ new Set(["approved", "changes_requested", "commented"]);
+    REVIEW_ACTION = "review_submitted";
     ID_CHARSET = /^[A-Za-z0-9._-]+$/;
     REPLICAS_MAX = 3;
+    INSTRUCTIONS_MAX = 2e3;
   }
 });
 
@@ -8442,7 +8510,7 @@ import { Queue } from "bullmq";
 function makeQueue(connection) {
   return new Queue(QUEUE, { connection });
 }
-async function enqueueLocalJob(queue, { folder, flow, task, provider, model, maxTurns, image, chainDepth, parentJobId, jobId, now = /* @__PURE__ */ new Date() }) {
+async function enqueueLocalJob(queue, { folder, flow, task, provider, model, maxTurns, image, skillsDir, chainDepth, parentJobId, jobId, now = /* @__PURE__ */ new Date() }) {
   const minute = now.toISOString().slice(0, 16);
   const id = jobId ?? localJobId({ folder, flow, task, minute });
   const data = {
@@ -8454,6 +8522,11 @@ async function enqueueLocalJob(queue, { folder, flow, task, provider, model, max
     model,
     maxTurns,
     ...image !== void 0 && { image },
+    // The host directory of operator-authored skills this trigger injects (REQ-PER-TRIGGER-SKILLS).
+    // Conditional like `image`, so an unflagged job's data stays byte-identical, and at JOB level rather
+    // than inside `trigger` because a worker-host path is an execution knob, not a fact about the
+    // delivery -- and `trigger` is the object copied into /job/event.json.
+    ...skillsDir !== void 0 && { skillsDir },
     ...chainDepth !== void 0 && { chainDepth },
     ...parentJobId !== void 0 && { parentJobId }
   };
@@ -8534,14 +8607,40 @@ var init_git_dirty = __esm({
   }
 });
 
+// ../worker/src/host-pi.mjs
+var init_host_pi = __esm({
+  "../worker/src/host-pi.mjs"() {
+    init_packages();
+  }
+});
+
+// ../worker/src/copy-tree.mjs
+var ENTRY_NAME_RE, INJECT_LIMITS;
+var init_copy_tree = __esm({
+  "../worker/src/copy-tree.mjs"() {
+    ENTRY_NAME_RE = /^[a-z0-9](?:[a-z0-9_.-]{0,62}[a-z0-9])?$/i;
+    INJECT_LIMITS = Object.freeze({
+      // The binding one, and its reason is NOT disk. Every loaded skill contributes a <name> and a
+      // description (spec-capped at 1024 chars) to the SYSTEM PROMPT of every job of that trigger, so
+      // this bounds the cached prefix rather than the filesystem. UNVERIFIED figure, in the sense
+      // ISOLATION_FLAGS' --pids-limit=512 is: a ceiling on absurdity, not a measured budget.
+      maxDirs: 64,
+      maxFiles: 512,
+      maxBytes: 4 << 20,
+      maxDepth: 8
+    });
+  }
+});
+
 // ../worker/src/import-pi.mjs
 import { execFile as execFile2 } from "node:child_process";
 import { promisify as promisify2 } from "node:util";
-var ENTRY_NAME_RE, execFileAsync;
+var execFileAsync;
 var init_import_pi = __esm({
   "../worker/src/import-pi.mjs"() {
     init_packages();
-    ENTRY_NAME_RE = /^[a-z0-9](?:[a-z0-9_.-]{0,62}[a-z0-9])?$/i;
+    init_host_pi();
+    init_copy_tree();
     execFileAsync = promisify2(execFile2);
   }
 });
@@ -8627,14 +8726,14 @@ function readStageManifest({ globalPiDir, readFile = readFileSync, fileExists =
       const { name, version, dir } = entry;
       if (!isNonEmptyString5(name) || !isNonEmptyString5(version) || !isNonEmptyString5(dir)) return null;
       if (!ENTRY_NAME_RE.test(dir) || dir.length > MAX_DIR_LENGTH) return null;
-      packages.push({ name, version, dir });
+      packages.push({ name, version, dir, from: entry.from === FROM_HOST ? FROM_HOST : FROM_DECLARED });
     }
     return { stagedAt: typeof parsed.stagedAt === "string" ? parsed.stagedAt : null, packages };
   } catch {
     return null;
   }
 }
-var PACKAGES_SUBDIR, STAGE_MANIFEST, CONTAINER_PACKAGES_ROOT, MAX_DIR_LENGTH;
+var PACKAGES_SUBDIR, STAGE_MANIFEST, CONTAINER_PACKAGES_ROOT, FROM_DECLARED, FROM_HOST, MAX_DIR_LENGTH;
 var init_packages = __esm({
   "../worker/src/packages.mjs"() {
     init_config2();
@@ -8643,6 +8742,8 @@ var init_packages = __esm({
     PACKAGES_SUBDIR = "packages";
     STAGE_MANIFEST = "packages.json";
     CONTAINER_PACKAGES_ROOT = `${CONTAINER_GLOBAL_PI_DIR}/${PACKAGES_SUBDIR}`;
+    FROM_DECLARED = "pi-packages";
+    FROM_HOST = "host";
     MAX_DIR_LENGTH = 64;
   }
 });
@@ -9056,6 +9157,8 @@ function normalizeTriggerForDisplay(entry) {
   const flow = typeof run.flow === "string" ? run.flow : null;
   const packages = run.packages !== false;
   const image = typeof run.image === "string" && run.image.trim() !== "" ? run.image : null;
+  const skillsDir = typeof run.skillsDir === "string" && run.skillsDir.trim() !== "" ? run.skillsDir : null;
+  const instructions = typeof run.instructions === "string" && run.instructions.trim() !== "";
   const resume = run.resume === true;
   const forge = typeof run.kind === "string" && run.kind.trim() !== "" ? run.kind : null;
   const replicas = Number.isInteger(run.replicas) && run.replicas > 1 ? run.replicas : null;
@@ -9072,12 +9175,14 @@ function normalizeTriggerForDisplay(entry) {
         model: typeof run.model === "string" ? run.model : null,
         packages,
         image,
+        skillsDir,
+        instructions,
         resume
       };
     case "label":
-      return { type: "label", any: normalizeSelector(on.any), all: normalizeSelector(on.all), none: normalizeSelector(on.none), flow, packages, image, resume, replicas, forge };
+      return { type: "label", any: normalizeSelector(on.any), all: normalizeSelector(on.all), none: normalizeSelector(on.none), flow, packages, image, skillsDir, instructions, resume, replicas, forge };
     case "comment":
-      return { type: "comment", phrase: typeof on.phrase === "string" ? on.phrase : null, flow, packages, image, resume, replicas, forge };
+      return { type: "comment", phrase: typeof on.phrase === "string" ? on.phrase : null, flow, packages, image, skillsDir, instructions, resume, replicas, forge };
     case "pull_request":
       return {
         type: "pull_request",
@@ -9088,6 +9193,8 @@ function normalizeTriggerForDisplay(entry) {
         flow,
         packages,
         image,
+        skillsDir,
+        instructions,
         resume,
         replicas,
         forge
@@ -35570,19 +35677,21 @@ function triggerLine(t) {
   const forge = t?.forge && t.forge !== "github" ? `  [${t.forge}]` : "";
   const pkgs = t?.packages === true ? "  [packages]" : "";
   const img = t?.image ? `  [image ${t.image}]` : "";
+  const skl = t?.skillsDir ? `  [skills ${String(t.skillsDir).split(/[\\/]/).filter(Boolean).pop()}]` : "";
+  const ins = t?.instructions === true ? "  [instructions]" : "";
   const res = t?.resume === true ? "  [resume]" : "";
   const rep = t?.replicas > 1 ? `  [x${t.replicas}]` : "";
   switch (t?.type) {
     case "cron":
-      return `cron  ${t.id ?? "-"}  ${t.pattern ?? "-"} \u2192 ${t.folder ?? "-"}/${flow}${forge}${pkgs}${img}${res}${rep}`;
+      return `cron  ${t.id ?? "-"}  ${t.pattern ?? "-"} \u2192 ${t.folder ?? "-"}/${flow}${forge}${pkgs}${img}${skl}${ins}${res}${rep}`;
     case "label":
-      return `label  ${ruleClauses(t) || "(no selector)"} \u2192 ${flow}${forge}${pkgs}${img}${res}${rep}`;
+      return `label  ${ruleClauses(t) || "(no selector)"} \u2192 ${flow}${forge}${pkgs}${img}${skl}${ins}${res}${rep}`;
     case "comment":
-      return `comment  "${t.phrase ?? "-"}" \u2192 ${flow}${forge}${pkgs}${img}${res}${rep}`;
+      return `comment  "${t.phrase ?? "-"}" \u2192 ${flow}${forge}${pkgs}${img}${skl}${ins}${res}${rep}`;
     case "pull_request": {
       const clauses = ruleClauses(t);
       const action = `action[${(t.action ?? []).join(",")}]`;
-      return `pull_request  ${action}${clauses ? ` ${clauses}` : ""} \u2192 ${flow}${forge}${pkgs}${img}${res}${rep}`;
+      return `pull_request  ${action}${clauses ? ` ${clauses}` : ""} \u2192 ${flow}${forge}${pkgs}${img}${skl}${ins}${res}${rep}`;
     }
     default:
       return "(unknown trigger)";
@@ -36968,7 +37077,9 @@ function trustModel(t) {
       ];
     default:
       return [
-        t?.type === "comment" ? "authorized by a collaborator comment (author_association) + HMAC" : "authorized by a collaborator's label + HMAC webhook + author gate",
+        // A review trigger reads the REVIEWER's author_association, not the PR author's (issue #66), so it
+        // cannot share the generic line: an operator reading "author gate" would picture the wrong person.
+        t?.type === "comment" ? "authorized by a collaborator comment (author_association) + HMAC" : t?.type === "pull_request" && (t?.action ?? []).includes("review_submitted") ? "authorized by the REVIEWER's author_association (not the PR author's) + HMAC" : "authorized by a collaborator's label + HMAC webhook + author gate",
         "dedup by X-GitHub-Delivery GUID (redelivery-safe)",
         where
       ];
@@ -37987,8 +38098,8 @@ var init_setup_wizard = __esm({
     init_deployment_pointer();
     init_read_model();
     init_index();
-    RUNTIME_VERSION = "0.1.2";
-    RECEIVER_VERSION = "0.1.1";
+    RUNTIME_VERSION = "0.2.0";
+    RECEIVER_VERSION = "0.2.0";
     RUNTIME_PKG = "@edgehero/pi-dispatch";
     RECEIVER_PKG = "@edgehero/pi-dispatch-receiver";
     NUDGE_MARKER_BASENAME = "pi-dispatch-setup.nudged";
@@ -38190,7 +38301,7 @@ function registerTools(pi) {
   pi.registerTool({
     name: "dispatch_trigger_add",
     label: "pi-dispatch add trigger",
-    description: "Adds a trigger to triggers.json and applies it live. The operator MUST approve a confirm dialog showing the entry; with no interactive operator it is refused. `flow` is the .pi/skills/<name> skill the job runs (its SKILL.md is the agent's instructions). `kind` = cron|label|comment|pull_request. cron (local) needs id, pattern, `folder` (absolute host path the job runs in), `flow`, and `task` (the prompt text handed to the agent), and may set optional model/provider/maxTurns for that schedule (omit = deployment default). label needs labels[]+flow; comment needs phrase+flow; pull_request needs action[] (+ optional labels[]) + flow. Webhook triggers take an optional `forge` = github (default) | gitlab, which also decides which action words pull_request accepts: github is labeled|opened|synchronize|reopened, gitlab is open|update|reopen|approved. For webhook triggers the repo and the task come from the triggering issue/PR event \u2014 set only the match + flow \u2014 and they run under the deployment default model.",
+    description: "Adds a trigger to triggers.json and applies it live. The operator MUST approve a confirm dialog showing the entry; with no interactive operator it is refused. `flow` is the .pi/skills/<name> skill the job runs (its SKILL.md is the agent's instructions). `kind` = cron|label|comment|pull_request. cron (local) needs id, pattern, `folder` (absolute host path the job runs in), `flow`, and `task` (the prompt text handed to the agent), and may set optional model/provider/maxTurns for that schedule (omit = deployment default). label needs labels[]+flow; comment needs phrase+flow; pull_request needs action[] (+ optional labels[]) + flow. Webhook triggers take an optional `forge` = github (default) | gitlab, which also decides which action words pull_request accepts: github is labeled|opened|synchronize|reopened|review_submitted, gitlab is open|update|reopen|approved. A github review_submitted trigger may also set reviewState[] (approved|changes_requested|commented) to narrow which verdicts fire; omitted, all three do. For webhook triggers the repo and the task come from the triggering issue/PR event \u2014 set only the match + flow \u2014 and they run under the deployment default model.",
     executionMode: "sequential",
     parameters: typebox_exports.Object({
       kind: typebox_exports.String(),
@@ -39086,7 +39197,7 @@ var init_index = __esm({
     ];
     FORGE_PROMPT = "forge \u2014 github, gitlab, forgejo or azure";
     PR_ACTION_VOCAB = {
-      github: { hint: "labeled opened synchronize reopened", dflt: "labeled" },
+      github: { hint: "labeled opened synchronize reopened review_submitted", dflt: "labeled" },
       gitlab: { hint: "open update reopen approved", dflt: "update" },
       forgejo: { hint: "label_updated opened synchronized reopened", dflt: "label_updated" },
       azure: { hint: "created updated", dflt: "updated" }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@edgehero/pi-dispatch-admin",
-  "version": "0.5.0",
+  "version": "0.6.0",
   "description": "Operator console (a pi extension) + skill for pi-dispatch: run the pi coding agent as a self-hosted service. A /dispatch TUI for the queue, spend caps, run history, editable GitHub, GitLab, Forgejo and Azure DevOps triggers (cron/label/comment/PR/MR/work item), and scheduled pause windows — plus AI-operable, human-confirmed controls. Runs against a live pi-dispatch deployment.",
   "keywords": [
     "pi-package",

package/skills/operate-pi-dispatch/SKILL.md

@@ -128,9 +128,14 @@ trigger is the same — the `on.type`, the `{any, all, none}` label predicate, `
 Two things are NOT the same, and both refuse at load rather than misbehaving quietly:
 
 - **`pull_request` actions are the forge's own words.** GitHub takes
-  `labeled | opened | synchronize | reopened`; GitLab takes `open | update | reopen | approved`. A word
+  `labeled | opened | synchronize | reopened | review_submitted`; GitLab takes
+  `open | update | reopen | approved`. A word
   from the wrong forge is refused when the file is written. It would not break anything otherwise — it
   would simply never match an event, and the trigger would look configured while doing nothing.
+  `review_submitted` is GitHub's `pull_request_review` event: it fires on **every** submitted review, so
+  add `reviewState` (`approved | changes_requested | commented`) to narrow which verdicts are worth paying
+  for. That field is github-only and legal only beside `review_submitted`; anywhere else it refuses at
+  load, because a narrowing that cannot apply reads as one that does.
 - **One `comment` trigger per forge.** Two GitHub comment triggers are refused; one GitHub and one GitLab
   are fine.