chore(release): auto-bump version to 0.54.7 — earned by #249 (fix(research): extract ALL text blocks from mimo events — progressive output con)

Promotion record: #249

.gitea/PULL_REQUEST_TEMPLATE.md

@@ -83,3 +83,9 @@ DEPENDENCIES (epic + issues). -->
 
 ## Plain-language review notes
 <!-- Explain what Nick should inspect, approve, reject, or question. -->
+
+## Research provenance
+<!-- If this PR's issue went through a research round: link the research file
+(research/issue-<N>-research.md) and state which synthesis decisions were
+adopted. If no research round: write "none" — the spec-gate will flag it. -->
+-->

.gitea/workflows/ci.yml

@@ -80,6 +80,10 @@ jobs:
           python3 -m venv .test-venv
           .test-venv/bin/pip install --disable-pip-version-check -r requirements-ci.txt
           .test-venv/bin/pip install --disable-pip-version-check -e . --no-deps
+      - name: Umbrella impact gate (propagation contract)
+        run: |
+          python3 scripts/umbrella_impact.py check \
+            --base "$(git merge-base origin/master HEAD)" --head HEAD
       - name: Test with coverage (global floor + changed-lines ratchet)
         run: |
           # #149 coverage ratchet — both floors live in .coveragerc; raise,

.gitea/workflows/research-stage-headless.yml

@@ -21,19 +21,28 @@ name: Research stage (headless)
 on:
   issues:
     types: [labeled]
+  workflow_dispatch:
+    inputs:
+      issue_number:
+        description: "Issue to research"
+        required: true
+      round:
+        description: "Research round number"
+        required: false
+        default: "1"
 
 permissions:
   issues: write
   contents: write
 
 concurrency:
-  group: research-stage-${{ gitea.event.issue.number }}
+  group: research-stage-${{ gitea.event.inputs.issue_number || gitea.event.issue.number }}
   cancel-in-progress: false
 
 jobs:
   research:
     runs-on: rune-ci-desktop
-    if: gitea.event.label.name == 'research-pending'
+    if: ${{ gitea.event.label.name == 'research-pending' || gitea.event_name == 'workflow_dispatch' }}
     steps:
       - name: Check out master (research reads the codebase, not a PR)
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
@@ -47,56 +56,76 @@ jobs:
           HOME: /var/lib/gitea-desktop-runner
           XDG_CONFIG_HOME: /home/nick/.config
           XDG_DATA_HOME: /var/lib/gitea-desktop-runner/.local-share-mimo
-          ISSUE_NUMBER: ${{ gitea.event.issue.number }}
-          ISSUE_TITLE: ${{ gitea.event.issue.title }}
-          ISSUE_BODY: ${{ gitea.event.issue.body }}
+          ISSUE_NUMBER: ${{ gitea.event.inputs.issue_number || gitea.event.issue.number }}
+          GITEA_TOKEN: ${{ secrets.ROBOT_TOKEN }}
+          SERVER: ${{ gitea.server_url }}
         run: |
-          # Prompt = governed repo template (prompts/research-round.md) with
-          # placeholders substituted. The template is the TUI-callback seam:
-          # future L3 (Rune TUI) owns/overrides it via reviewed diffs.
+          # Fetch issue title/body from the API (dispatch events carry no issue
+          # payload). Deterministic, no scraping.
           python3 - <<'PY'
-          import os
+          import json, os, urllib.request
+          token = os.environ["GITEA_TOKEN"]
+          base = os.environ["SERVER"].rstrip("/") + "/api/v1/repos/conductor/rune"
+          n = os.environ["ISSUE_NUMBER"]
+          req = urllib.request.Request(
+              f"{base}/issues/{n}",
+              headers={"Authorization": "token " + token})
+          with urllib.request.urlopen(req, timeout=30) as r:
+              issue = json.loads(r.read())
           tpl = open("prompts/research-round.md").read()
-          out = (tpl.replace("{{ISSUE_NUMBER}}", os.environ["ISSUE_NUMBER"])
-                    .replace("{{ISSUE_TITLE}}", os.environ["ISSUE_TITLE"])
-                    .replace("{{ISSUE_BODY}}", os.environ["ISSUE_BODY"] or "(empty body)"))
+          out = (tpl.replace("{{ISSUE_NUMBER}}", n)
+                    .replace("{{ISSUE_TITLE}}", issue.get("title", ""))
+                    .replace("{{ISSUE_BODY}}", issue.get("body") or "(empty body)"))
           open("/tmp/research_prompt.txt", "w").write(out)
-          print(f"prompt built: {len(out)} chars from prompts/research-round.md")
+          print(f"prompt built: {len(out)} chars (issue #{n}: {issue.get('title','')[:50]})")
           PY
 
           /var/lib/gitea-desktop-runner/bin/mimo run \
+            --agent build \
             -m "Verboo Code/glm-5.3-flash" \
             --dangerously-skip-permissions \
             --format json \
             "$(cat /tmp/research_prompt.txt)" \
-            > /tmp/research_events.jsonl 2>/tmp/research_err.log
+            > /tmp/research_events.jsonl 2>/tmp/research_err.log || {
+            echo "mimo exited nonzero — stderr:"
+            cat /tmp/research_err.log | tail -10
+            exit 1
+          }
+          echo "mimo events: $(wc -l < /tmp/research_events.jsonl) lines"
 
-          # Extract final assistant text (fail-loud: no swallow)
+          # Extract ALL text parts (the agent may produce multiple text blocks
+          # across tool-use rounds; the last contiguous block is the synthesis)
           python3 - <<'PY'
           import json
-          text = ""
+          texts = []
           try:
               for line in open("/tmp/research_events.jsonl"):
                   try:
                       ev = json.loads(line)
                   except json.JSONDecodeError:
                       continue
-                  if ev.get("type") == "assistant" and isinstance(ev.get("message", {}).get("content"), list):
-                      for part in ev["message"]["content"]:
-                          if part.get("type") == "text":
-                              text = part.get("text", text)
+                  if ev.get("type") == "text":
+                      t = ev.get("part", {}).get("text", "")
+                      if t.strip():
+                          texts.append(t)
           except FileNotFoundError:
               pass
-          if not text.strip():
-              raise SystemExit("research stage: agent produced NO text — infrastructure or model failure (fail-loud)")
+          if not texts:
+              raise SystemExit(
+                  f"research stage: agent produced NO text in {sum(1 for _ in open('/tmp/research_events.jsonl'))} events "
+                  "— infrastructure or model failure (fail-loud)"
+              )
+          # Join all text blocks (the agent's progressive output IS the round)
+          text = "\n\n".join(texts)
           open("/tmp/research_round.md", "w").write(text)
+          print(f"research text: {len(text)} chars from {len(texts)} blocks")
           PY
 
       - name: File research as comment + versioned research file
         env:
           GITEA_TOKEN: ${{ secrets.ROBOT_TOKEN }}
           SERVER: ${{ gitea.server_url }}
-          ISSUE_NUMBER: ${{ gitea.event.issue.number }}
+          ISSUE_NUMBER: ${{ gitea.event.inputs.issue_number || gitea.event.issue.number }}
         run: |
           python3 - <<'PY'
           import json, os, subprocess, urllib.request
@@ -119,6 +148,14 @@ jobs:
           with urllib.request.urlopen(req, timeout=30) as r:
               print(f"research comment posted: {r.status}")
 
+          # Disposition routing (§14): PROCEED → research-clear fires via the
+          # marker comment. MORE_RESEARCH → strip the synthesis from the
+          # comment is NOT needed (the gate keys on all 4 phase markers, which
+          # are present); instead the run_agent_stage result records the
+          # disposition and the ROUND increments on the next dispatch. The
+          # scheduled sweep re-dispatches when research-pending persists and
+          # no stage is running (§19 reconciliation path).
+
           # 2. Research file committed straight to master by the robot identity
           subprocess.run(["git", "config", "user.name", "conductor"], check=True)
           subprocess.run(["git", "config", "user.email", "nick@thingscouldgetdicey.online"], check=True)

README.md

@@ -1,6 +1,6 @@
 # Rune
 
-**Current release: v0.54.1**
+**Current release: v0.54.7**
 
 **Rune is a persistent personal agent designed to accumulate experience, develop reusable competence, and remain recognisably the same agent even when its underlying foundation model changes.**
 

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rune"
-version = "0.54.1"
+version = "0.54.7"
 description = "Rune — a stateful coding assistant with persistent memory, orchestration, and TUI"
 requires-python = ">=3.10"
 

scripts/research_artifact.py

@@ -0,0 +1,97 @@
+#!/usr/bin/env python3
+"""research_artifact.py — structured research-round schema + validation (P0.4).
+
+Per the Forge Cognitive Research Loop handoff §8: machine-readable research
+round with causal evidence links. Phases are FROZEN artifacts — the later
+phase's input is the persisted prior artifact (§7: "the later phase must not
+be able to pretend it saw research that the pipeline never recorded").
+
+Schema (minimal, causally sufficient):
+  issue, round, receipt (from run_agent_stage.build_receipt),
+  phases: {inward: {claims: [{id, claim, evidence[]}]},
+           outward: {sources: [{id, url, title, retrieved_at, tool,
+                                content_hash, learned}],
+                     claims: [{id, claim, supports[]}]},
+           innovation: {candidates: [{id, proposal, derived_from[],
+                                      expected_gain, risks[]}]},
+           synthesis: {decision, evidence_used[], blocking_unknowns[],
+                       feasibility, kill_test,
+                       disposition: PROCEED|MORE_RESEARCH}}
+"""
+
+from __future__ import annotations
+
+import json
+import re
+import sys
+from pathlib import Path
+
+
+def validate_outward(outward: dict) -> list[str]:
+    """§9: OUTWARD valid only with real discovery — ≥2 sources, each with
+    url + tool receipt + learned note; ≥1 search query recorded."""
+    errs = []
+    sources = outward.get("sources", [])
+    if len(sources) < 2:
+        errs.append(f"OUTWARD: {len(sources)} sources < 2 (real discovery required)")
+    for s_ in sources:
+        for field in ("id", "url", "tool", "learned"):
+            if not s_.get(field):
+                errs.append(f"OUTWARD source {s_.get('id', '?')}: missing {field}")
+        if not re.match(r"https?://", s_.get("url", "")):
+            errs.append(f"OUTWARD source {s_.get('id', '?')}: url not http(s)")
+    if not outward.get("queries"):
+        errs.append("OUTWARD: no search queries recorded (discovery, not recall)")
+    return errs
+
+
+def validate_causal_links(artifact: dict) -> list[str]:
+    """§11/§12: innovation candidates derive_from prior claims; synthesis
+    decisions cite evidence_used. At least one candidate depends on OUTWARD
+    evidence unless synthesis records 'no external mechanism relevant'."""
+    errs = []
+    phases = artifact.get("phases", {})
+    inward_ids = {c["id"] for c in phases.get("inward", {}).get("claims", [])}
+    outward_ids = {c["id"] for c in phases.get("outward", {}).get("claims", [])}
+    candidates = phases.get("innovation", {}).get("candidates", [])
+    synthesis = phases.get("synthesis", {})
+    known = inward_ids | outward_ids
+    for c in candidates:
+        refs = set(c.get("derived_from", []))
+        if not refs:
+            errs.append(f"INNOVATION {c.get('id', '?')}: no derived_from (decorative, not causal)")
+            continue
+        if not refs & known:
+            errs.append(f"INNOVATION {c.get('id', '?')}: derived_from references unknown claims {sorted(refs - known)}")
+    outward_dependent = any(set(c.get("derived_from", [])) & outward_ids for c in candidates)
+    if not outward_dependent and "no external mechanism relevant" not in synthesis.get("decision", ""):
+        errs.append("no INNOVATION candidate derives from OUTWARD evidence — research was decorative")
+    if not synthesis.get("evidence_used"):
+        errs.append("SYNTHESIS: evidence_used empty — decisions untraceable")
+    if synthesis.get("disposition") == "MORE_RESEARCH" and not synthesis.get("blocking_unknowns"):
+        errs.append("MORE_RESEARCH disposition without named blocking_unknowns")
+    return errs
+
+
+def validate(artifact: dict) -> list[str]:
+    errs = validate_outward(artifact.get("phases", {}).get("outward", {}))
+    errs += validate_causal_links(artifact)
+    if artifact.get("phases", {}).get("synthesis", {}).get("disposition") not in ("PROCEED", "MORE_RESEARCH"):
+        errs.append("synthesis.disposition must be PROCEED or MORE_RESEARCH")
+    return errs
+
+
+def main_with(argv: list[str]) -> int:
+    artifact = json.loads(Path(argv[1]).read_text())
+    errs = validate(artifact)
+    if errs:
+        print("research artifact INVALID:")
+        for e in errs:
+            print(f"  - {e}")
+        return 1
+    print("research artifact VALID (outward discovery + causal links verified)")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main_with(sys.argv))

scripts/research_backfill_sweep.py

@@ -133,8 +133,13 @@ def compute_targets(issues: list[dict], api) -> list[dict]:
         if unsatisfied:
             continue
         _, comments = api.get(f"/issues/{i['number']}/comments")
-        if has_research_round(comments if isinstance(comments, list) else []):
-            continue
+        comments = comments if isinstance(comments, list) else []
+        if has_research_round(comments):
+            # A round exists — re-queue ONLY if its disposition says MORE_RESEARCH
+            # (§14: the loop continues). Latest round wins (comments are append-only).
+            more = any(re.search(r"Disposition:\s*MORE_RESEARCH", c.get("body", ""), re.IGNORECASE) for c in comments)
+            if not more:
+                continue
         targets.append(i)
     return targets
 

src/rune/__init__.py

@@ -3,4 +3,4 @@
 This is the main entry point for the Rune program.
 """
 
-__version__ = "0.54.1"
+__version__ = "0.54.7"

src/rune/canary_gitea_automation.py

@@ -252,7 +252,7 @@ def main() -> int:
     )
     check(
         "Rune minor version advertises provisioning and restore verification",
-        '__version__ = "0.54.1"' in rune_version,
+        '__version__ = "0.54.7"' in rune_version,
         failures,
     )
     # #169 policy update: auto-bump on merge is sanctioned (Nick directive:
@@ -270,7 +270,7 @@ def main() -> int:
     )
     check(
         "release job explicitly binds its Gitea token",
-        "GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }}" in versioning,
+        "GITEA_TOKEN: ${{ secrets.ROBOT_TOKEN }}" in versioning,
         failures,
     )
 

src/rune/rune_tui.py

@@ -7,7 +7,7 @@ mode system, subject header, shared memory and identity.
 
 This is not an interface. This is how I see, speak, and interact with the world.
 
-Version: v0.54.1
+Version: v0.54.7
 Features: 41 (24 original + 17 new native commands)
 
 Usage:
@@ -1923,7 +1923,7 @@ class RuneApp(App):
         # Show version info
         self.chat_channel.add_message(
             "system",
-            f"[dim]Rune v0.54.1 | Textual TUI | {len(loaded)} native modules[/dim]",
+            f"[dim]Rune v0.54.7 | Textual TUI | {len(loaded)} native modules[/dim]",
         )
 
         # Auto-show night mode suggestion if evening/night hours

tests/test_research_artifact.py

@@ -0,0 +1,227 @@
+"""Red-first tests for scripts/research_artifact.py — P0.4 causal validation.
+
+Per the Forge Cognitive Research Loop handoff:
+  §9  OUTWARD = real discovery (≥2 sources with url+tool+learned, ≥1 query)
+  §11 INNOVATION candidates MUST derive_from prior claims; ≥1 from OUTWARD
+  §12 SYNTHESIS decisions cite evidence_used
+  §14 MORE_RESEARCH requires named blocking_unknowns
+"""
+
+from __future__ import annotations
+
+import importlib.util
+from pathlib import Path
+
+import pytest
+
+ROOT = Path(__file__).resolve().parent.parent
+SPEC = importlib.util.spec_from_file_location("research_artifact", ROOT / "scripts" / "research_artifact.py")
+ra = importlib.util.module_from_spec(SPEC)
+SPEC.loader.exec_module(ra)
+
+
+def make_artifact(**overrides):
+    base = {
+        "issue": 42,
+        "round": 1,
+        "receipt": {"search_or_fetch_calls": 3, "distinct_web_domains": 2},
+        "phases": {
+            "inward": {
+                "claims": [{"id": "I1", "claim": "store uses FTS5", "evidence": ["src/rune/memory_store.py:86"]}]
+            },
+            "outward": {
+                "queries": ["python sqlite fts5 hybrid search"],
+                "sources": [
+                    {
+                        "id": "O1",
+                        "url": "https://sqlite.org/fts5.html",
+                        "title": "FTS5",
+                        "retrieved_at": "2026-09-08T00:00Z",
+                        "tool": "webfetch",
+                        "content_hash": "abc",
+                        "learned": "FTS5 supports external content tables",
+                    },
+                    {
+                        "id": "O2",
+                        "url": "https://github.com/asg017/sqlite-vec",
+                        "title": "sqlite-vec",
+                        "retrieved_at": "2026-09-08T00:00Z",
+                        "tool": "webfetch",
+                        "content_hash": "def",
+                        "learned": "vec0 virtual tables for embedding search",
+                    },
+                ],
+                "claims": [{"id": "O-C1", "claim": "hybrid FTS+vec is viable", "supports": ["O1", "O2"]}],
+            },
+            "innovation": {
+                "candidates": [
+                    {
+                        "id": "C1",
+                        "proposal": "hybrid recall via FTS+vec union",
+                        "derived_from": ["I1", "O-C1"],
+                        "expected_gain": "better recall",
+                        "risks": ["latency"],
+                    },
+                ]
+            },
+            "synthesis": {
+                "decision": "adopt hybrid recall",
+                "evidence_used": ["I1", "O-C1", "C1"],
+                "blocking_unknowns": [],
+                "feasibility": "mechanism proven by O2",
+                "kill_test": "recall returns >0 rows on seeded store",
+                "disposition": "PROCEED",
+            },
+        },
+    }
+    for k, v in overrides.items():
+        if k == "phases":
+            base["phases"].update(v)
+        else:
+            base[k] = v
+    return base
+
+
+def test_valid_artifact_passes():
+    assert ra.validate(make_artifact()) == []
+
+
+def test_outward_single_source_fails():
+    a = make_artifact(
+        phases={
+            "outward": {
+                "queries": ["q"],
+                "sources": [{"id": "O1", "url": "https://x", "tool": "webfetch", "learned": "y"}],
+                "claims": [],
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("sources < 2" in e for e in errs)
+
+
+def test_outward_missing_learned_fails():
+    a = make_artifact(
+        phases={
+            "outward": {
+                "queries": ["q"],
+                "sources": [
+                    {"id": "O1", "url": "https://x", "tool": "webfetch", "learned": ""},
+                    {"id": "O2", "url": "https://y", "tool": "webfetch", "learned": "z"},
+                ],
+                "claims": [],
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("missing learned" in e for e in errs)
+
+
+def test_outward_no_queries_fails():
+    a = make_artifact(
+        phases={
+            "outward": {
+                "queries": [],
+                "sources": [
+                    {"id": "O1", "url": "https://x", "tool": "webfetch", "learned": "y"},
+                    {"id": "O2", "url": "https://y", "tool": "webfetch", "learned": "z"},
+                ],
+                "claims": [],
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("no search queries" in e for e in errs)
+
+
+def test_innovation_without_derived_from_fails():
+    a = make_artifact(
+        phases={
+            "innovation": {
+                "candidates": [{"id": "C1", "proposal": "x", "derived_from": [], "expected_gain": "", "risks": []}]
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("no derived_from" in e for e in errs)
+
+
+def test_innovation_unknown_claim_reference_fails():
+    a = make_artifact(
+        phases={
+            "innovation": {
+                "candidates": [
+                    {"id": "C1", "proposal": "x", "derived_from": ["I999"], "expected_gain": "", "risks": []}
+                ]
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("unknown claims" in e for e in errs)
+
+
+def test_no_outward_derived_candidate_fails_unless_exempt():
+    a = make_artifact(
+        phases={
+            "innovation": {
+                "candidates": [{"id": "C1", "proposal": "x", "derived_from": ["I1"], "expected_gain": "", "risks": []}]
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("decorative" in e for e in errs)
+    # exemption: synthesis explicitly records no external mechanism relevant
+    a["phases"]["synthesis"]["decision"] = "internal-only; no external mechanism relevant"
+    errs2 = ra.validate(a)
+    assert not any("decorative" in e for e in errs2)
+
+
+def test_synthesis_empty_evidence_used_fails():
+    a = make_artifact(
+        phases={
+            "synthesis": {
+                "decision": "x",
+                "evidence_used": [],
+                "blocking_unknowns": [],
+                "feasibility": "",
+                "kill_test": "",
+                "disposition": "PROCEED",
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("evidence_used empty" in e for e in errs)
+
+
+def test_more_research_without_unknowns_fails():
+    a = make_artifact(
+        phases={
+            "synthesis": {
+                "decision": "x",
+                "evidence_used": ["I1"],
+                "blocking_unknowns": [],
+                "feasibility": "",
+                "kill_test": "",
+                "disposition": "MORE_RESEARCH",
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("blocking_unknowns" in e for e in errs)
+
+
+def test_invalid_disposition_fails():
+    a = make_artifact(
+        phases={
+            "synthesis": {
+                "decision": "x",
+                "evidence_used": ["I1"],
+                "blocking_unknowns": [],
+                "feasibility": "",
+                "kill_test": "",
+                "disposition": "VIBES",
+            }
+        }
+    )
+    errs = ra.validate(a)
+    assert any("disposition" in e for e in errs)

tests/test_research_sweep.py

@@ -174,3 +174,19 @@ def test_skips_issues_with_open_pr(forge, monkeypatch):
     monkeypatch.setattr(rs, "open_pr_issue_refs", lambda api: {30})
     targets = rs.compute_targets(forge.issues, forge)
     assert [t["number"] for t in targets] == [31]
+
+
+def test_more_research_round_not_treated_as_done(forge):
+    """P0.6 §14: a completed round with MORE_RESEARCH disposition must NOT be
+    treated as done (the backfill sweep must not skip it as 'already
+    researched' if the label was lost — the stage re-dispatches)."""
+    forge.issues = [_issue(32)]  # label lost (event delivery failure)
+    forge.comments = {32: [{"body": RESEARCH_COMMENT + "\n\nDisposition: MORE_RESEARCH (blocking: U1 perf numbers)"}]}
+    targets = rs.compute_targets(forge.issues, forge)
+    assert [t["number"] for t in targets] == [32], "MORE_RESEARCH stays eligible for re-dispatch"
+
+
+def test_proceed_round_not_requeued(forge):
+    forge.issues = [_issue(33)]  # NOT labeled (research-clear already ran)
+    forge.comments = {33: [{"body": RESEARCH_COMMENT + "\n\nDisposition: PROCEED"}]}
+    assert rs.compute_targets(forge.issues, forge) == []