@exaudeus/workrail 3.70.1 → 3.70.2

package/dist/manifest.json

@@ -473,8 +473,8 @@
       "sha256": "5fe866e54f796975dec5d8ba9983aefd86074db212d3fccd64eed04bc9f0b3da",
       "bytes": 8011
     },
-    "console-ui/assets/index-BcZJOyVG.js": {
-      "sha256": "30e50a4bf35f3383569bf5c19e27ab255611f2524048ce306017e88ddffb2b69",
+    "console-ui/assets/index-Cr14LfsQ.js": {
+      "sha256": "d5ab351690b555194e079cdfe4bedcf22ac8644568d9a2b5a622d07f3451d291",
       "bytes": 767983
     },
     "console-ui/assets/index-DHrKiMCf.css": {
@@ -482,7 +482,7 @@
       "bytes": 60673
     },
     "console-ui/index.html": {
-      "sha256": "813832d8d8ed9de1597a4c972111a165117cfaa97674a0bbb66783f8de954cb8",
+      "sha256": "c79189d614e1dacfc75ab3a9c4e1e030677f8730e8c2423ac45f977abd010c1a",
       "bytes": 417
     },
     "console/standalone-console.d.ts": {
@@ -614,8 +614,8 @@
       "bytes": 1216
     },
     "daemon/daemon-events.d.ts": {
-      "sha256": "ba8fbd7be2cdd5f0f53c45b3ed4f27b51739ca3cb782e931629469597f98250d",
-      "bytes": 5359
+      "sha256": "469cc9b6954e19a5eb87c3fac42f96fd66cd31321b33ce741fb9c1aa64cb2b80",
+      "bytes": 5369
     },
     "daemon/daemon-events.js": {
       "sha256": "b6841eef4634bb266faf81961c1e387b535dd64a74d58582f3f2bad8c3469d95",
@@ -658,8 +658,8 @@
       "bytes": 8385
     },
     "daemon/workflow-runner.js": {
-      "sha256": "0575f958423e4f0a4abbc712e0901c5bee8c5ed5f2ffa4b3a68083f4251e1243",
-      "bytes": 102945
+      "sha256": "87e8dfbc87f2794f79b4c591eb65a204f12761e3b438174da0f81097fd1e5f4a",
+      "bytes": 103183
     },
     "di/container.d.ts": {
       "sha256": "003bb7fb7478d627524b9b1e76bd0a963a243794a687ff233b96dc0e33a06d9f",
@@ -1718,12 +1718,12 @@
       "bytes": 6968
     },
     "trigger/polling-scheduler.d.ts": {
-      "sha256": "60df456a31fa87ce71de76f5e31a6c460bfab588a24c8a2f06bf926fdcea550a",
-      "bytes": 1096
+      "sha256": "c8bcd28794a23906feabe276725e63fe7af79749e0517a46c6e0ed41da0cabb2",
+      "bytes": 1152
     },
     "trigger/polling-scheduler.js": {
-      "sha256": "a546506d1586b0020a64f1321c07d929f7b24920a0fbd18d6566ff73fb6d9185",
-      "bytes": 22409
+      "sha256": "0f8264ac6393d91e22cc48d6fb89095de2391420175009b68b5848c13be3f7f0",
+      "bytes": 22481
     },
     "trigger/trigger-listener.d.ts": {
       "sha256": "cbd89c24cdfe89cb555946b0dcaa6836bf81f37db56c5e80d408dc4a2fe42fb9",
@@ -2066,8 +2066,8 @@
       "bytes": 1664
     },
     "v2/durable-core/domain/prompt-renderer.js": {
-      "sha256": "fe7d177a79cf8cbf2ee814a0e20e24b4cb3ea1b332c4f6aaf57d0f601a95419b",
-      "bytes": 22976
+      "sha256": "caff32c39ccd59219b44a5063e00e73dd29e5ba3c86a0602c2e609d17e8a0105",
+      "bytes": 23600
     },
     "v2/durable-core/domain/reason-model.d.ts": {
       "sha256": "a944e7e0d9b3c73468488263cb0aa1e446c023f8084fd2af53cbda3f3bfcd37a",

package/dist/trigger/polling-scheduler.d.ts

@@ -16,10 +16,11 @@ export declare class PollingScheduler {
     private readonly router;
     private readonly store;
     private readonly fetchFn?;
+    private readonly sessionsDir;
     private readonly intervals;
     private readonly polling;
     private readonly dispatchingIssues;
-    constructor(triggers: readonly TriggerDefinition[], router: TriggerRouter, store: PolledEventStore, fetchFn?: FetchFn | undefined);
+    constructor(triggers: readonly TriggerDefinition[], router: TriggerRouter, store: PolledEventStore, fetchFn?: FetchFn | undefined, sessionsDir?: string);
     start(): void;
     stop(): void;
     forcePoll(triggerId: string): Promise<ForcePollResult>;

package/dist/trigger/polling-scheduler.js

@@ -46,11 +46,12 @@ function isPollingTrigger(trigger) {
     return trigger.pollingSource !== undefined;
 }
 class PollingScheduler {
-    constructor(triggers, router, store, fetchFn) {
+    constructor(triggers, router, store, fetchFn, sessionsDir = path.join(os.homedir(), '.workrail', 'daemon-sessions')) {
         this.triggers = triggers;
         this.router = router;
         this.store = store;
         this.fetchFn = fetchFn;
+        this.sessionsDir = sessionsDir;
         this.intervals = new Map();
         this.polling = new Map();
         this.dispatchingIssues = new Set();
@@ -222,7 +223,7 @@ class PollingScheduler {
             await appendQueuePollLog({ event: 'poll_cycle_complete', triggerId, reason: 'not_implemented', queueType: queueConfig.type, ts: new Date().toISOString() });
             return;
         }
-        const sessionsDir = path.join(os.homedir(), '.workrail', 'daemon-sessions');
+        const sessionsDir = this.sessionsDir;
         const activeSessions = await countActiveSessions(sessionsDir);
         if (activeSessions >= queueConfig.maxTotalConcurrentSessions) {
             console.log(`[QueuePoll] Skipping cycle: active sessions (${activeSessions}) >= maxTotalConcurrentSessions (${queueConfig.maxTotalConcurrentSessions}).`);

package/dist/v2/durable-core/domain/prompt-renderer.js

@@ -216,16 +216,16 @@ function buildMetricsSection(profile, isLastStep, cleanFormat) {
             if (!isLastStep)
                 return shaFooter;
             const finalFooter = cleanFormat
-                ? '\n\nMetrics (final): also set metrics_outcome, metrics_pr_numbers, metrics_files_changed, metrics_lines_added, metrics_lines_removed in context.'
-                : '\n\n**METRICS (System):** This is the final step. Also report:\n- `metrics_outcome`: `"success"` | `"partial"` | `"abandoned"` | `"error"`\n- `metrics_pr_numbers`: array of integer PR numbers (not URLs)\n- `metrics_files_changed`: integer count\n- `metrics_lines_added`: integer count\n- `metrics_lines_removed`: integer count\n\nCall `continue_workflow` with all of the above in `context: { metrics_commit_shas: [...], metrics_outcome: "success", ... }`.';
+                ? '\n\nMetrics (final): also set metrics_outcome (exactly one of: "success", "partial", "abandoned", "error"), metrics_pr_numbers, metrics_files_changed, metrics_lines_added, metrics_lines_removed in context.'
+                : '\n\n**METRICS (System):** This is the final step. Also report:\n- `metrics_outcome`: set to exactly one of these four strings -- no other values are valid: `"success"`, `"partial"`, `"abandoned"`, `"error"`. Do not describe what you did -- classify the outcome using only these values.\n- `metrics_pr_numbers`: array of integer PR numbers (not URLs)\n- `metrics_files_changed`: integer count\n- `metrics_lines_added`: integer count\n- `metrics_lines_removed`: integer count\n\nCall `continue_workflow` with all of the above in `context: { metrics_commit_shas: [...], metrics_outcome: "success", ... }`.';
             return shaFooter + finalFooter;
         }
         case 'review': {
             if (!isLastStep)
                 return '';
             return cleanFormat
-                ? '\n\nMetrics (final): set metrics_pr_numbers (integer array) and metrics_outcome in context.'
-                : '\n\n**METRICS (System):** This is the final step of a review workflow. Report:\n- `metrics_pr_numbers`: array of integer PR numbers reviewed (not URLs)\n- `metrics_outcome`: `"success"` | `"partial"` | `"abandoned"` | `"error"`\n\nCall `continue_workflow` with `context: { metrics_pr_numbers: [123], metrics_outcome: "success" }`.';
+                ? '\n\nMetrics (final): set metrics_pr_numbers (integer array) and metrics_outcome (exactly one of: "success", "partial", "abandoned", "error") in context.'
+                : '\n\n**METRICS (System):** This is the final step of a review workflow. Report:\n- `metrics_pr_numbers`: array of integer PR numbers reviewed (not URLs)\n- `metrics_outcome`: set to exactly one of these four strings -- no other values are valid: `"success"`, `"partial"`, `"abandoned"`, `"error"`. Do not describe what you did -- classify the outcome using only these values.\n\nCall `continue_workflow` with `context: { metrics_pr_numbers: [123], metrics_outcome: "success" }`.';
         }
         case 'research':
         case 'design':
@@ -233,8 +233,8 @@ function buildMetricsSection(profile, isLastStep, cleanFormat) {
             if (!isLastStep)
                 return '';
             return cleanFormat
-                ? '\n\nMetrics (final): set metrics_outcome in context.'
-                : '\n\n**METRICS (System):** This is the final step. Report:\n- `metrics_outcome`: `"success"` | `"partial"` | `"abandoned"` | `"error"`\n\nCall `continue_workflow` with `context: { metrics_outcome: "success" }`.';
+                ? '\n\nMetrics (final): set metrics_outcome (exactly one of: "success", "partial", "abandoned", "error") in context.'
+                : '\n\n**METRICS (System):** This is the final step. Report:\n- `metrics_outcome`: set to exactly one of these four strings -- no other values are valid: `"success"`, `"partial"`, `"abandoned"`, `"error"`. Do not describe what you did -- classify the outcome using only these values.\n\nCall `continue_workflow` with `context: { metrics_outcome: "success" }`.';
         }
     }
 }

package/docs/discovery/design-review-findings.md

@@ -1,110 +1,107 @@
-# Design Review Findings: wr.discovery Goal Reframing
+# Design Review Findings: Issue #393 Stale Tracker Close
 
-*Concise, actionable findings for main-agent synthesis. Not a final decision.*
-**Date:** 2026-04-18
+**Reviewed design:** Candidate A -- Close Issue #393 with evidence comment  
+**Review date:** 2026-04-23  
+**Reviewer:** wr.discovery session (design_first / QUICK path)
 
 ---
 
 ## Tradeoff Review
 
-**Tradeoff 1: goalType classification remains agent judgment**
+**Tradeoff 1: Close comment mentions why auto-close did not fire**
+- Does not violate any decision criterion -- in fact satisfies Criterion 5 (captures the learning)
+- Tone risk (sounds like blame): mitigated by using technical/mechanical language
+- Hidden assumption: maintainer reads close comments -- reasonable for a 1-person repo
+- **Verdict: Acceptable**
 
-- Acceptable: this pattern is established in the workflow (rigorMode, pathRecommendation are all agent-derived)
-- Classification examples in the procedure reduce misclassification probability
-- **Finding: YELLOW.** Add goalType classification examples to procedure text to reduce ambiguity at the problem_framed / opportunity_framed boundary.
-
-**Tradeoff 2: overhead for well-framed goals is nonzero**
-
-- A few additional context variable captures and procedure lines in Phase 0
-- Well-framed goals produce minimal output ('goalType = problem_framed, no impliedProblem needed')
-- **Finding: NON-ISSUE.** Overhead is trivial.
-
-**Tradeoff 3: Phase 1g always-on for design_first/full_spectrum**
-
-- One additional advance per session for these paths
-- Produces trivially short output for well-framed sessions ('pathChangedAfterContext = false')
-- **CRITICAL CORRECTION NEEDED:** Phase 1g runCondition must be an OR (`retriageNeeded = true OR pathRecommendation in [design_first, full_spectrum]`) not a replacement. Otherwise landscape_first sessions that explicitly need retriage will not trigger it.
-- **Finding: YELLOW.** Runnable as-designed if the OR condition is used correctly.
+**Tradeoff 2: No AGENTS.md update -- pattern prevention sacrificed**
+- Actively satisfies Criterion 3 (no new problems) by not touching a protected human-maintained file
+- Only one observed instance of the failure mode (PR #790) -- insufficient evidence of a recurring pattern
+- The learning is captured in the close comment itself, which is more contextually located than AGENTS.md
+- **Verdict: Acceptable**
 
 ---
 
 ## Failure Mode Review
 
-**Failure mode 1: Agent misclassifies solution-framed goal as opportunity_framed**
-- Status: **Partially mitigated.** C3's Phase 1e/1f required 'what would make this framing wrong' output provides a downstream catch. Classification examples in Phase 0 reduce probability.
-- Missing mitigation: examples in procedure (address in revisions)
-- **Finding: MEDIUM risk, mitigated.**
+**FM1: Maintainer re-opens because they wanted to close personally**
+- Design handles it: close comment provides full evidence chain; re-open is 5-second CLI command; no data loss
+- Likelihood: Low (0 comments, 2-day stale, daemon assignee, no checkbox activity)
+- Severity if occurs: Minimal (issue re-opens, maintainer closes manually, traceability comment persists)
+- Missing mitigations: None needed
 
-**Failure mode 2: 'What would make this framing wrong' output is formulaic**
-- Status: **Partially mitigated.** Making it required non-empty enforces form but not quality.
-- Missing mitigation: specificity instruction ('name ONE concrete condition, not a general caveat')
-- **Finding: LOW-MEDIUM risk.**
-
-**Failure mode 3: Phase 1g doesn't surface new insights for well-framed sessions**
-- Status: **Non-issue by design.** For well-framed sessions, Phase 1g is a graceful no-op that confirms the path is still correct. One advance wasted, nothing more.
-- **Finding: LOW risk, acceptable.**
+**FM2: CI secretly failing for the test file**
+- Design handles it: 14/14 passing verified locally immediately before action; test is isolated unit coverage
+- No related CI failure issues exist in the open issues list for this test file
+- Severity if occurs: Low (tracker close is orthogonal to CI state; no test regression introduced)
+- Missing mitigations: None needed
 
 ---
 
 ## Runner-Up / Simpler Alternative Review
 
-**C2 (mandatory Phase 0a):** The structural enforcement advantage is real but comes at the cost of a mandatory overhead step for all sessions. The C1+C3 hybrid achieves most of C2's value via procedure-level enforcement plus structural runCondition changes. C2 is the right escalation if the hybrid proves insufficient.
-
-**Simpler variant:** Just one sentence added to Phase 0: 'If the goal is solution-framed, derive the underlying problem.' Too narrow -- no context variables means no downstream reference to the reframing.
-
-**`alternativeFraming` addition from C2:** Borrowing C2's `alternativeFraming` requirement (one reframe even when the original goal seems correct) is high-value, low-cost. Add to Phase 0 design doc entry, not as a context variable.
+**Runner-up (Candidate B -- close + AGENTS.md note):**
+- Candidate B's only distinct value (keyword failure mode documentation) is fully absorbed into Candidate A's close comment
+- No element of B is orphaned; no hybrid needed
 
-**Finding: C1+C3 hybrid with two refinements (examples, alternativeFraming) stands. No direction change needed.**
+**Simpler variant (close without comment):**
+- Fails Criterion 2 (no traceability) and Criterion 5 (no learning capture)
+- Not viable -- the comment is load-bearing, not decorative
 
 ---
 
 ## Philosophy Alignment
 
-| Principle | Status |
-|---|---|
-| Validate at boundaries, trust inside | SATISFIED -- Phase 0 becomes an active validator |
-| Make illegal states unrepresentable | PARTIALLY SATISFIED -- C3 structural changes help; C2 would fully satisfy |
-| YAGNI with discipline | SATISFIED -- no new steps, minimal additions |
-| Architectural fixes over patches | SATISFIED -- runCondition changes and required output contracts are structural |
-| Determinism over cleverness | SATISFIED -- same goalType input produces same path behavior |
+**Clearly satisfied:**
+- Validate at boundaries, trust inside -- all validation done before action
+- Observability -- close comment makes the state transition and rationale fully visible
+- Document "why" not "what" -- comment explains rationale, not just action
+- Atomicity -- single CLI call, no partial state possible
+- Architectural fixes over patches -- reframe correctly identified the real problem (stale tracker) before acting
 
-**One explicit philosophy tension:** 'Make illegal states unrepresentable' vs 'YAGNI with discipline' -- deliberately accepted, C2 is escalation path.
+**Under tension:**
+- Agent authority over human-filed issues -- mild tension; resolved by reversibility + transparent comment
+- **Verdict: Acceptable tension, not risky**
 
 ---
 
 ## Findings
 
-### Yellow findings
+No RED or ORANGE findings. All challenges and reviews converge.
 
-**Y1: goalType classification boundary ambiguity**
-The boundary between `problem_framed` and `opportunity_framed` is unclear without examples. Add classification examples to Phase 0 procedure to reduce misclassification at this boundary.
-
-**Y2: Phase 1g runCondition must be OR, not replacement**
-The retriage step runCondition must be: `retriageNeeded = true OR pathRecommendation == design_first OR pathRecommendation == full_spectrum`. A straight replacement would break landscape_first sessions that legitimately need retriage.
-
-**Y3: 'What would make this framing wrong' needs specificity instruction**
-The required output field should specify 'name ONE concrete falsification condition, not a general caveat.' Without this, the field can be satisfied by formulaic responses.
-
-### No Red or Orange findings
-
-The selected C1+C3 direction has no material structural weaknesses.
+**YELLOW -- Tone of close comment (INFO)**
+- Risk: the explanation of why auto-close did not fire could read as attributing a mistake to the PR author
+- Mitigation: use mechanical/technical language ("GitHub requires `Closes #NNN` syntax; PR #790 used different phrasing") rather than evaluative language
+- Action: word the comment accordingly -- no structural change to the design needed
 
 ---
 
 ## Recommended Revisions
 
-1. **Add goalType classification examples** to Phase 0 procedure (solution_framed: 'add X', 'implement Y', 'build X'; problem_framed: 'reduce X', 'fix Y'; opportunity_framed: 'explore X', 'decide whether Y'; decision_framed: 'choose between A and B')
+**Revision 1 (from YELLOW finding): Prescribe exact comment wording**
 
-2. **Add `alternativeFraming`** as a required design doc entry in Phase 0: 'Before selecting a path, generate one alternative framing -- if the stated goal is wrong, what would a better goal be?'
+Use this comment text:
 
-3. **Use OR condition for Phase 1g runCondition:** `retriageNeeded = true OR pathRecommendation in [design_first, full_spectrum]`
+> All acceptance criteria for this issue are satisfied on `main`.
+>
+> - `loadSessionNotes` is exported at `src/daemon/workflow-runner.ts` (added in PR #790)
+> - All 4 failure paths (token decode, store load, projection, unexpected exception) and the happy path are covered by `tests/unit/workflow-runner-load-session-notes.test.ts` (added in PR #782)
+> - 14 tests pass: `npx vitest run tests/unit/workflow-runner-load-session-notes.test.ts`
+>
+> Note: PR #790 referenced this issue as "Closes issue #393 pre-existing test failures" but GitHub's auto-close requires the exact syntax `Closes #393` -- the non-standard phrasing is why the issue was not automatically closed on merge.
 
-4. **Add specificity instruction** to Phase 1e/1f 'what would make this framing wrong' field: require naming one concrete falsification condition.
+This wording is factual, neutral, and gives any future reader everything they need to verify or re-open.
 
 ---
 
 ## Residual Concerns
 
-1. The goalType classification is LLM-dependent. Without empirical testing on real sessions, we cannot confirm the classification is reliable. This is an inherent limitation of the approach.
+**RC1 (low): Pattern recurrence unaddressed**
+If the PR keyword failure mode recurs on a second PR, the case for Candidate B (AGENTS.md note) strengthens. This is not actionable now but should be noted for future monitoring.
+
+**RC2 (very low): Open CI failures on main**
+There are 10+ open "CI failure on main blocking release" issues. These are unrelated to the test file in question (verified locally). If main CI is broken in a way that affects this test file, the close would be slightly premature. Probability is very low given local verification.
+
+---
 
-2. The C1+C3 hybrid does not prevent path-selection bias for the window between Phase 0 path selection and Phase 1g retriage. A session that selects the wrong path in Phase 0 runs several steps in the wrong direction before Phase 1g can correct it. Acceptable for STANDARD rigor; C2 is the correct escalation if this proves problematic.
+**Overall verdict: PROCEED with Candidate A as designed, using the prescribed comment wording from Revision 1.**