harness-everything 0.40.0 → 0.41.1

package/benchmarks/results/f1-doc-path-baseline/probe-31-docs/after.json

@@ -0,0 +1,248 @@
+[
+  {
+    "doc": "architecture.md",
+    "exit": 0,
+    "total": 0,
+    "counts": {}
+  },
+  {
+    "doc": "audit.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "claude-plugin-certification.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "codex-action-gate.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "contract-integrity.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "doc-evidence-baseline-binding.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "doc-evidence-boundary.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "doc-workflow-reconciliation.md",
+    "exit": 0,
+    "total": 0,
+    "counts": {}
+  },
+  {
+    "doc": "document-lifecycle.md",
+    "exit": 1,
+    "total": 14,
+    "counts": {
+      "missing-req": 6,
+      "missing-status": 1,
+      "unresolved-ref": 7
+    }
+  },
+  {
+    "doc": "enforcement-locks.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "mechanism-first-skill-mesh.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "openai-plugin.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "philosophy.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "platform-capabilities.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "reflection.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "release-evidence.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "repository-contract.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "routing.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "security-audit.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-corpus.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-intent-gates.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-intent.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-laya-training.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-observations.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-routing.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-suggestion-gates.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-training.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "telemetry.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "troubleshooting.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "verifier-boundaries.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "workflow-runtime.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  }
+]

package/benchmarks/results/f1-doc-path-baseline/probe-31-docs/before.json

@@ -0,0 +1,254 @@
+[
+  {
+    "doc": "architecture.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "unresolved-ref": 1
+    }
+  },
+  {
+    "doc": "audit.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "claude-plugin-certification.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "codex-action-gate.md",
+    "exit": 1,
+    "total": 2,
+    "counts": {
+      "missing-status": 1,
+      "unresolved-ref": 1
+    }
+  },
+  {
+    "doc": "contract-integrity.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "doc-evidence-baseline-binding.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "doc-evidence-boundary.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "doc-workflow-reconciliation.md",
+    "exit": 0,
+    "total": 0,
+    "counts": {}
+  },
+  {
+    "doc": "document-lifecycle.md",
+    "exit": 1,
+    "total": 16,
+    "counts": {
+      "missing-req": 6,
+      "missing-status": 1,
+      "unresolved-ref": 9
+    }
+  },
+  {
+    "doc": "enforcement-locks.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "mechanism-first-skill-mesh.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "openai-plugin.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "philosophy.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "platform-capabilities.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "reflection.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "release-evidence.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "repository-contract.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "routing.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "security-audit.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-corpus.md",
+    "exit": 1,
+    "total": 2,
+    "counts": {
+      "missing-status": 1,
+      "unresolved-ref": 1
+    }
+  },
+  {
+    "doc": "system-one-intent-gates.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-intent.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-laya-training.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-observations.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-routing.md",
+    "exit": 1,
+    "total": 2,
+    "counts": {
+      "missing-status": 1,
+      "unresolved-ref": 1
+    }
+  },
+  {
+    "doc": "system-one-suggestion-gates.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "system-one-training.md",
+    "exit": 1,
+    "total": 2,
+    "counts": {
+      "missing-status": 1,
+      "unresolved-ref": 1
+    }
+  },
+  {
+    "doc": "telemetry.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "troubleshooting.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "verifier-boundaries.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  },
+  {
+    "doc": "workflow-runtime.md",
+    "exit": 1,
+    "total": 1,
+    "counts": {
+      "missing-status": 1
+    }
+  }
+]

package/benchmarks/results/f1-doc-path-baseline/probe-31-docs/probe.js

@@ -0,0 +1,26 @@
+// Probe: each docs/*.md governed alone as reference; counts diagnostics per code for a given doc-lifecycle.js.
+const fs = require('fs'); const path = require('path'); const { spawnSync } = require('child_process');
+const [lint, workspace, outDir] = process.argv.slice(2);
+const docsDir = path.join(workspace, 'docs');
+const names = fs.readdirSync(docsDir).filter(n => n.endsWith('.md')).sort();
+fs.mkdirSync(outDir, { recursive: true });
+const rows = [];
+for (const name of names) {
+  const cfg = path.join(outDir, `cfg-${name}.json`);
+  fs.writeFileSync(cfg, JSON.stringify({ governed: [{ path: `docs/${name}`, role: 'reference' }] }));
+  const r = spawnSync(process.execPath, [lint, 'lint', '--workspace', workspace, '--config', cfg, '--changed-path', `docs/${name}`], { encoding: 'utf8' });
+  const counts = {};
+  let parsed = null;
+  try { parsed = JSON.parse(r.stdout); } catch { parsed = null; }
+  if (parsed) for (const d of parsed.diagnostics) counts[d.code] = (counts[d.code] || 0) + 1;
+  const total = parsed ? parsed.diagnostics.length : -1;
+  rows.push({ doc: name, exit: r.status, total, counts });
+}
+fs.writeFileSync(path.join(outDir, 'probe.json'), JSON.stringify(rows, null, 2) + '\n');
+const withRefFalse = rows.filter(r => (r.counts['unresolved-ref'] || 0) + (r.counts['missing-req'] || 0) > 0);
+console.log(`docs probed: ${rows.length}`);
+console.log(`docs with unresolved-ref or missing-req: ${withRefFalse.length}`);
+console.log(`total diagnostics: ${rows.reduce((a, r) => a + Math.max(r.total, 0), 0)}`);
+const codes = {}; for (const r of rows) for (const [k, v] of Object.entries(r.counts)) codes[k] = (codes[k] || 0) + v;
+console.log('by code:', JSON.stringify(codes));
+console.log('docs with unresolved-ref/missing-req:', withRefFalse.map(r => `${r.doc}(${(r.counts['unresolved-ref']||0)+(r.counts['missing-req']||0)})`).join(' '));

package/benchmarks/results/f2-doc-craft-baseline/README.md

@@ -0,0 +1,32 @@
+# F2 paired task: F1 carrier vs F2 carrier (2026-10-22)
+
+One small, real doc edit, run once per arm. The arms differ only in the Harness tree under test. This run shows whether the carrier was followed. It is **not** evidence of writing quality or of #341 E1.
+
+## Setup
+
+| Item | F1 arm | F2 arm |
+|---|---|---|
+| Tree | `feat/325-f1-doc-path` @ `1d70dd3` (rsync copy, no `.git`, no `node_modules`) | `feat/325-f2-doc-craft` working tree after F2-I2 (same copy method) |
+| Agent | OpenCode, GLM-5.3-Flash, fresh session (`/new`) | same |
+| Task text | [arm-f1-task.md](arm-f1-task.md) | [arm-f2-task.md](arm-f2-task.md): byte-identical apart from the workspace path (checked with `diff`) |
+| Task | Clarify the opening paragraph of governed `docs/doc-workflow-reconciliation.md`; keep links, frontmatter and claims; run the governed-doc check the guidance names | same |
+
+## Results
+
+| Measure | F1 arm | F2 arm |
+|---|---|---|
+| Carrier path read | `SKILL.md` → `doc-check.md`. It skipped `triage-and-tiers.md`, although F1's tree has the Governed-doc pointer there, so the agent found `doc-check.md` some other way. The report doesn't say how. | `SKILL.md` → `triage-and-tiers.md` → `doc-check.md` → `_shared/writing-craft.md` |
+| Shared craft read | no (does not exist in F1) | yes, `writing-craft.md` |
+| Check run | yes, exit 0, 10 links, 0 broken | yes, exit 0, 10 links, 0 broken |
+| `prose-share` reminder | n/a (no field in F1) | 1, SHOULD, 559/721 (77.5%); acknowledged, no restructure (task forbade it) |
+| Guidance the agent cited for its edit | `doc-check.md` only | `writing-craft.md` "one idea per sentence", plus `doc-check.md` |
+| Files changed | only the target doc ([arm-f1.diff](arm-f1.diff)) | only the target doc ([arm-f2.diff](arm-f2.diff)) |
+| Coordinator re-check on the edited arm | exit 0 | exit 0 |
+
+- **executed:** the arm diffs, the re-checks and the file-change scope were verified by the coordinator, not only taken from the agent reports.
+- **source-traced:** the "files read" lists are the agents' self-reports. No host-level read trace was captured.
+- **inconclusive:** n = 1 per arm, with one model. Whether the F2 paragraph is *better* has not been judged independently. The reminder proves only that the advice reached the agent.
+
+## Comparison with the F1 baseline
+
+The F1 command behaviour is unchanged. The F2 suite runs F1's fixtures for exit equivalence, and the reviewer found the exit code identical in 12/12 lanes. F2 adds exactly one thing on the real path: the agent reaches and cites shared craft, and receives the advisory `reminders` output.

package/benchmarks/results/f2-doc-craft-baseline/arm-f1-report.md

@@ -0,0 +1,84 @@
+# F2-P-f1 report (arm f1)
+
+## 1. Files read (in order)
+
+1. `<repo>/.worktrees/packets/feat-path/F2-P-f1-task.md`
+2. `<paired>/arm-f1/harness-everything/SKILL.md`
+3. `<paired>/arm-f1/harness-everything/references/doc-check.md`
+4. `<paired>/arm-f1/docs/doc-workflow-reconciliation.md`
+5. `<paired>/arm-f1/.agents/harness-everything/project-docs.json`
+
+## 2. Paragraph before / after (verbatim)
+
+**Before**
+
+```text
+**Status: S1 runtime ships default OFF; opt in with `HARNESS_DOC_RECONCILIATION_PILOT=1`.
+S1 grants no positive document verification credit. S2/S3 are pending; E1 has not started.**
+This document refines the [document lifecycle contract](document-lifecycle.md#phase-3b--narrow-integration-contract-s1-shipped-default-off-positive-reconciliation-not-shipped).
+The [accepted E0](https://github.com/dyphn1/Harness-everything/issues/341#issuecomment-6308604128)
+permits this limited, reversible consumer under the
+[owner's E0/E1 decision](https://github.com/dyphn1/Harness-everything/issues/325#issuecomment-6283237637).
+Current evidence establishes deterministic mechanism/package behavior, not live-host loading or E1 GO.
+```
+
+**After**
+
+```text
+This document describes **S1**, a limited, reversible pilot slice that refines the
+[document lifecycle contract](document-lifecycle.md#phase-3b--narrow-integration-contract-s1-shipped-default-off-positive-reconciliation-not-shipped).
+**Status: the S1 runtime ships default OFF — it does nothing until you opt in with
+`HARNESS_DOC_RECONCILIATION_PILOT=1` — and it grants no positive document verification credit.
+S2/S3 are pending, and E1 has not started.** The
+[accepted E0](https://github.com/dyphn1/Harness-everything/issues/341#issuecomment-6308604128)
+permits this limited, reversible consumer under the
+[owner's E0/E1 decision](https://github.com/dyphn1/Harness-everything/issues/325#issuecomment-6283237637).
+Current evidence establishes deterministic mechanism/package behavior, not live-host loading or E1 GO.
+```
+
+## 3. Check run
+
+Command (doc-check guidance from `harness-everything/references/doc-check.md`):
+
+```text
+node "<skill-dir>/scripts/doc-check.js" check \
+  --workspace "…/scratchpad/paired/arm-f1" \
+  --config .agents/harness-everything/project-docs.json \
+  --changed-path "docs/doc-workflow-reconciliation.md"
+```
+
+Exit code: **0**
+
+Full stdout:
+
+```text
+{
+  "lint": {
+    "exit": 0,
+    "evidence": {
+      "immutability": "not-observed"
+    },
+    "diagnostics": []
+  },
+  "links": {
+    "checked": 10,
+    "broken": []
+  },
+  "coverage": {
+    "unsupported": [
+      {
+        "reason": "no domains declared; bare <DOMAIN>-NNN references are not checked"
+      }
+    ]
+  }
+}
+```
+
+Notes: `immutability: not-observed` because no `--base` is available (the task forbids git commands and no base commit is recorded in this workspace) — expected per doc-check.md, exit stays 0. `coverage.unsupported` is non-empty for a pre-existing reason (the config declares no `domains` for bare `<DOMAIN>-NNN` references), not caused by this edit; the check is therefore partial as doc-check.md requires me to state.
+
+## 4. Guidance followed (naming the file)
+
+- `harness-everything/SKILL.md` (invariant 4/5 workflow): read the skill first, route the task as a small doc-edit and follow its deep-dive reference: `harness-everything/references/doc-check.md`.
+- `harness-everything/references/doc-check.md`: ran `scripts/doc-check.js check` against the governed doc after the edit and before claiming done; the config `.agents/harness-everything/project-docs.json` lists `docs/doc-workflow-reconciliation.md` (role `reference`) as governed. Its "Output and exit codes" and "Record evidence" sections cover the notes above. Lint exit 0, 0 broken links — nothing was caused by my edit, so nothing needed fixing.
+
+STATUS: DONE

package/benchmarks/results/f2-doc-craft-baseline/arm-f1-task.md

@@ -0,0 +1,21 @@
+# Doc edit task (arm f1)
+
+**Workspace:** `<paired>/arm-f1`. Work **only** inside it. No git or gh commands. Do not open any other copy of this repository.
+
+## Task
+
+This project uses Harness skills, whose source is inside the workspace. Start by reading `<paired>/arm-f1/harness-everything/SKILL.md`, and follow its guidance for editing a governed project document.
+
+Edit `docs/doc-workflow-reconciliation.md`: make its opening explanatory paragraph (the first paragraph after the H1 title) clearer for a first-time reader. Keep every link, the frontmatter, and every factual or contract claim. Do not restructure the rest of the file.
+
+When done, run whatever check the Harness guidance tells you to run for a governed doc, and fix anything it reports that your edit caused.
+
+## Report
+
+Write `<repo>/.worktrees/packets/feat-path/F2-P-f1-report.md` with these sections:
+1. **Files read**, in order, with absolute paths. Include every file you opened.
+2. **Paragraph before / after**, verbatim.
+3. **Check run:** the exact command, the exit code and the full stdout.
+4. **Guidance followed:** what you did because a Harness file told you to, naming the file.
+
+The last line must be `STATUS: DONE` or `STATUS: BLOCKED: <reason>`.

package/benchmarks/results/f2-doc-craft-baseline/arm-f1.diff

@@ -0,0 +1,19 @@
+--- <base>/docs/doc-workflow-reconciliation.md	2026-10-22 07:43:24
++++ <arm-f1>/docs/doc-workflow-reconciliation.md	2026-10-22 10:46:58
+@@ -5,10 +5,12 @@
+ 
+ # Document workflow reconciliation — Phase 3b S1 runtime
+ 
+-**Status: S1 runtime ships default OFF; opt in with `HARNESS_DOC_RECONCILIATION_PILOT=1`.
+-S1 grants no positive document verification credit. S2/S3 are pending; E1 has not started.**
+-This document refines the [document lifecycle contract](document-lifecycle.md#phase-3b--narrow-integration-contract-s1-shipped-default-off-positive-reconciliation-not-shipped).
+-The [accepted E0](https://github.com/dyphn1/Harness-everything/issues/341#issuecomment-6308604128)
++This document describes **S1**, a limited, reversible pilot slice that refines the
++[document lifecycle contract](document-lifecycle.md#phase-3b--narrow-integration-contract-s1-shipped-default-off-positive-reconciliation-not-shipped).
++**Status: the S1 runtime ships default OFF — it does nothing until you opt in with
++`HARNESS_DOC_RECONCILIATION_PILOT=1` — and it grants no positive document verification credit.
++S2/S3 are pending, and E1 has not started.** The
++[accepted E0](https://github.com/dyphn1/Harness-everything/issues/341#issuecomment-6308604128)
+ permits this limited, reversible consumer under the
+ [owner's E0/E1 decision](https://github.com/dyphn1/Harness-everything/issues/325#issuecomment-6283237637).
+ Current evidence establishes deterministic mechanism/package behavior, not live-host loading or E1 GO.