factoryline-code-factory 0.44.2__tar.gz → 0.44.3__tar.gz

factoryline_code_factory-0.44.3/.zenodo.json

@@ -0,0 +1,83 @@
+{
+  "title": "Code Factory: Proof-by-Sabotage Software Factory",
+  "description": "Code Factory answers one practical question for developers using AI: can this test actually fail? Run `factory first-proof --root .` to see a safe negative-control demonstration and receive a local, reviewable receipt. For teams, `factory wrap` records an admitted agent run's exact file delta, runs declared independent validators, and stores hashes and bounded facts rather than prompts or raw model output. Platform and assurance teams can evaluate policy gates, evidence packets, expiring exceptions, and tenant boundaries in a controlled pilot; this beta does not claim a support SLA, compliance certification, customer references, or procurement readiness. Version 0.44.3 adds Journey Reality (declared-versus-observed journey comparison), bounded failure capsules, stateful workflow checks, proof-gated healing, and independently audits the agent proposing a repair. Advanced Unified Graph Ops views can export bounded proof paths as Mermaid diagrams for offline review; optional deeper checks include LangGraph resume parity and a Gauntlet Survival Card.\n\nEvidence boundary: First Proof is a disposable demonstration, not an assessment of the user's project. Proof Cards summarize one verified receipt and exclude commands, paths, repository names, prompts, logs, and identity. They do not certify production readiness, security, coverage, identity, or release authority. The 60-day personal-use case is illustrative—not a benchmark, guaranteed ROI, or verified cash saving—and net savings must subtract tool cost and human oversight.\n\nCase-study visual: https://raw.githubusercontent.com/zrk222/code-factory/v0.44.3/docs/assets/marketplace/code-factory-60-day-personal-case-study.png\n\nDual licensed under MIT or Apache-2.0.",
+  "version": "0.44.3",
+  "publication_date": "2026-08-25",
+  "creators": [
+    {
+      "name": "Katz, Richard"
+    }
+  ],
+  "access_right": "open",
+  "license": "MIT",
+  "upload_type": "software",
+  "keywords": [
+    "software-factory",
+    "ai-agents",
+    "workflow-verification",
+    "mutation-testing",
+    "release-receipts",
+    "developer-tools",
+    "target-compiler",
+    "capability-packs",
+    "pack-composition",
+    "mcp-server",
+    "dsse-ed25519",
+    "product-missions",
+    "signal-loop",
+    "opinion-dock",
+    "independent-verification",
+    "product-graph",
+    "vscode-extension",
+    "jetbrains-plugin",
+    "postgresql-rls",
+    "github-app",
+    "oidc",
+    "transactional-outbox",
+    "tenant-control-plane",
+    "operator-console",
+    "langgraph",
+    "byok-routing",
+    "multi-provider-routing",
+    "human-in-the-loop",
+    "product-captures",
+    "content-addressed-proof",
+    "proof-reuse",
+    "intake-grill",
+    "proof-delta-loop",
+    "gauntlet",
+    "survival-card",
+    "graph-ops",
+    "unified-graph",
+    "prd-grill",
+    "verifier-plane",
+    "github-pull-request",
+    "proof-review",
+    "plan-to-proof-review",
+    "proof-debt",
+    "design-review",
+    "ui-quality",
+    "e2e-proof",
+    "team-pilot",
+    "developer-memory",
+    "live-activity",
+    "langgraph-assurance",
+    "resume-parity",
+    "agent-replay",
+    "github-actions",
+    "mcp-registry",
+    "local-stdio",
+    "journey-reality",
+    "failure-capsule",
+    "stateful-workflow-proof",
+    "proof-gated-healing",
+    "agent-work-audit"
+  ],
+  "related_identifiers": [
+    {
+      "identifier": "https://github.com/zrk222/code-factory",
+      "relation": "isSupplementTo",
+      "scheme": "url"
+    }
+  ]
+}

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/CITATION.cff

@@ -7,5 +7,5 @@ authors:
     given-names: Richard
 repository-code: "https://github.com/zrk222/code-factory"
 license: "MIT OR Apache-2.0"
-version: 0.44.2
-date-released: 2026-08-21
+version: 0.44.3
+date-released: 2026-08-25

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: factoryline-code-factory
-Version: 0.44.2
-Summary: Catch AI-generated tests that could never fail and review AI code with local proof.
+Version: 0.44.3
+Summary: Catch AI-generated tests that could never fail with one local proof command.
 Author: Richard Katz
 Author-email: rkatz22@gmail.com
 Maintainer: Richard Katz
@@ -60,22 +60,38 @@ Dynamic: license-file
 
 > **Catch AI-generated tests that could never fail — before review.**
 
-> **Free, local proof for code built with AI.** Start from a plain-language
-> outcome, a fuzzy PRD, or a risky diff. Code Factory binds the declared intent,
-> challenges whether a test can actually reject the failure it claims to cover,
-> and shows the current proof gap plus the next human decision. Tests that only
-> look green are not proof: a starter is never called production-ready before
-> the relevant proof exists.
+> **One command answers one question: can this test actually fail?** Code Factory
+> runs a safe negative control, records what happened, and gives you a local
+> result you can review. It challenges whether a test can actually reject the
+> failure it claims to cover. It does not upload your code or call a green test
+> proof when that test could never catch the failure.
 
 ```powershell
 pip install factoryline-code-factory
 factory first-proof --root .
 ```
 
-That one local command creates a disposable sandbox, proves the healthy check
-can pass, then proves the hollow check is caught as `HOLLOW_E2E_TEST`. It writes
-a receipt and a privacy-safe Proof Card; it never uploads your code. When you
-are ready to build, run `factory mvp "Build an approval tracker" --root .`.
+That one local command creates a disposable demonstration sandbox, proves the
+healthy check can pass, then proves the hollow check is rejected. It writes a
+receipt and a privacy-safe Proof Card; it does not assess or change your project.
+
+### Pick the path that matches your work
+
+- **Individual developer or vibe coder:** start with `factory first-proof`.
+  Learn what a real negative control looks like before trusting AI-generated tests.
+- **Engineering team:** put `factory wrap` around an admitted Codex, Claude Code,
+  DeepSeek, or other local agent run. Code Factory records the exact file delta,
+  runs the validators your team declared, and stores hashes and bounded facts—not
+  prompts or raw model output—so reviewers can compare the claim with the change.
+- **Platform or assurance team:** evaluate local policy gates, evidence packets,
+  expiring exceptions, and tenant boundaries in a controlled pilot. Code Factory
+  is beta software; it does not provide a support SLA, compliance certification,
+  customer references, or procurement readiness.
+
+When you need more than the first proof, open the [advanced workflow guide](docs/OVERVIEW.md).
+To create a contained starter later, run
+`factory mvp "Build an approval tracker" --root .`.
+That starter is never called production-ready until the relevant proof exists.
 
 ![Actual privacy-safe First Proof Card: the hollow test was detected](docs/assets/first-proof-card.svg)
 
@@ -90,7 +106,7 @@ are ready to build, run `factory mvp "Build an approval tracker" --root .`.
 
 ![Actual Factory Studio showing the outcome-first local MVP path and local-only boundary](docs/assets/marketplace/factory-studio-mvp-1280x800.png)
 
-## What it does
+## Advanced workflows
 
 - **Start a real project.** Turn one outcome into a contained web, mobile, API,
   CLI, worker, MCP, or agent-UI starting state.
@@ -126,6 +142,13 @@ are ready to build, run `factory mvp "Build an approval tracker" --root .`.
   review date, and hash-bound proof obligations. A Change Safety Case routes
   one explicit diff to its owner; it never infers intent, promotes a decision,
   runs a test, or approves a release.
+- **Prove the whole journey—and audit the repairing agent.** Use
+  [`factory journey`](docs/JOURNEY_PROOF_ENGINE.md) to compare declared and
+  observed product journeys, preserve a hash-bound failure capsule, prove
+  state flow and cleanup, and challenge a proposed healing with a negative
+  mutation. Choose human-controlled or bounded supervised-auto mode; every
+  autonomous attempt receives a separate FactoryLine audit and neither mode
+  can self-approve.
 
 ## What the latest upgrade resolves
 
@@ -137,6 +160,7 @@ are ready to build, run `factory mvp "Build an approval tracker" --root .`.
 | Teams gradually trust an agent because it has been successful recently | Earned Autonomy derives an expiring local tier from governed evidence and demotes severe hollow-test or scope-escape results | Identity, permissions, execution, approval, and release authority |
 | A hard-won design decision is forgotten, then a later diff silently breaks its assumptions | Engineering Judgment Capsules bind one named owner, explicit path scope, review date, and declared proof obligations; an optional human-declared Change Profile makes novel boundaries and required Senior Attention explicit before review | Proposal, independent promotion, decision reconsideration, proof execution, acceptance, merge, and release |
 | A reviewer has suggestions but no shared evidence picture | Graph Ops, local receipts, and read-only MCP facts show current scope, proof debt, and the next fact-derived action | Merge, release, deployment, and provider access |
+| A selector repair makes the test green but changes the user journey—or an autonomous agent silently edits outside scope | Journey Reality, Stateful Workflow Proof, and Proof-Gated Healing require preserved semantic identity and coverage, cleanup proof, an adversarial negative result, exact workspace deltas, and an independent Agent Work Audit | Agent/provider choice, repair admission, final approval, merge, and release |
 | Agent work happens outside the evidence ledger, while writing real E2E manifests takes too long | Evidence Supply Line wraps any admitted local agent CLI; `gauntlet draft` proposes inert, structure-derived promise drafts and explicitly withholds commands it cannot derive | Agent identity, sandboxing, draft promotion, validator choice, Gauntlet admission, and release |
 | A large/remote workspace feels opaque or sluggish | Workspace Advisor measures bounded local project shape and path-only WSL/remote signals without changing indexes, heap, caches, inspections, or remote settings | Every IDE performance change and environment setting |
 
@@ -147,11 +171,11 @@ security, productivity, production readiness, or an automatic repair service.
 
 | Who is using it | Start here | Highest-value uses | The result they can inspect |
 | --- | --- | --- | --- |
-| **Novice / first-time builder** | `factory first-proof`, then `factory mvp` | Catch a test that only looks green; turn one outcome into a contained starter; learn the difference between generated code and proven behavior | A privacy-safe Proof Card, a local MVP, and a visible next step instead of an unexplained pass/fail |
+| **Individual developer / vibe coder** | `factory first-proof` | See whether a passing test can actually reject a known failure before trusting it | A local receipt and privacy-safe Proof Card from a disposable demonstration; your project remains unchanged |
 | **Junior developer** | `factory prd grill`, `factory plan verify`, `factory change review` | Clarify acceptance evidence before coding; keep an AI-assisted diff inside the approved plan; surface missing tests and Proof Debt before review | Source-bound questions, exact changed paths, severity-ordered findings, and a review handoff |
 | **Senior / staff engineer** | `factory judgment`, `factory graph forensics`, `factory proofsearch`, `factory gauntlet` | Protect architecture decisions; diagnose resumed or parallel workflow drift; compare candidate repairs; challenge whether critical E2E checks can actually reject declared failures | Hash-bound decision context, first-divergence facts, rejected candidates, a deterministic winner, and Survival Cards |
-| **Engineering team** | `factory wrap`, GitHub Proof Review, Graph Ops, Agent License and Combine | Put agent work into one evidence ledger; add a neutral PR Check beside AI review; hand off proof state; compare sealed runs without turning scores into vendor claims | Commit-bound Checks, local receipts, explicit owners and blockers, proof reuse decisions, and comparable governed-run records |
-| **Enterprise / platform team** | Intake Grill, Verifier Plane, policy challenge, assurance dossier, admission controls | Standardize approval boundaries; test whether policy rules matter; keep signers, exceptions, tenants, budgets, and release evidence explicit; integrate with existing SDLC controls | Independently verifiable packets and read-only operational views; identity, credentials, merge, release, and deployment remain in enterprise-owned systems |
+| **Engineering team** | `factory wrap`, then GitHub Proof Review | Record the exact agent-created file delta, run declared independent validators, and hand reviewers a commit-bound evidence packet without retaining prompts | Local receipts, exact changed paths, explicit blockers, and a neutral PR Check beside AI review |
+| **Platform / assurance pilot** | Policy gates and evidence packets | Evaluate named approval boundaries, expiring exceptions, tenant isolation, budgets, and release evidence inside existing SDLC controls | Independently verifiable packets and read-only views; identity, credentials, merge, release, and deployment remain enterprise-owned. No SLA, certification, references, or procurement-readiness claim |
 
 ## Expected operational outcomes
 
@@ -386,7 +410,7 @@ context reusable by a client you choose.
   optional design-quality lane and its explicit review boundaries.
 - Read [The approval signal decays when AI-written code becomes routine](docs/HABITUATION_ESSAY.md)
   for the design and limits of the habituation gate.
-- See the [release notes](docs/RELEASE_NOTES_0.44.2.md),
+- See the [release notes](docs/RELEASE_NOTES_0.44.3.md),
   [CHANGELOG.md](CHANGELOG.md), [release channels](docs/RELEASE_CHANNELS.md), and
   [publication guide](PUBLICATION_GUIDE.md) for versioned release detail.
 

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/README.md

@@ -11,22 +11,38 @@
 
 > **Catch AI-generated tests that could never fail — before review.**
 
-> **Free, local proof for code built with AI.** Start from a plain-language
-> outcome, a fuzzy PRD, or a risky diff. Code Factory binds the declared intent,
-> challenges whether a test can actually reject the failure it claims to cover,
-> and shows the current proof gap plus the next human decision. Tests that only
-> look green are not proof: a starter is never called production-ready before
-> the relevant proof exists.
+> **One command answers one question: can this test actually fail?** Code Factory
+> runs a safe negative control, records what happened, and gives you a local
+> result you can review. It challenges whether a test can actually reject the
+> failure it claims to cover. It does not upload your code or call a green test
+> proof when that test could never catch the failure.
 
 ```powershell
 pip install factoryline-code-factory
 factory first-proof --root .
 ```
 
-That one local command creates a disposable sandbox, proves the healthy check
-can pass, then proves the hollow check is caught as `HOLLOW_E2E_TEST`. It writes
-a receipt and a privacy-safe Proof Card; it never uploads your code. When you
-are ready to build, run `factory mvp "Build an approval tracker" --root .`.
+That one local command creates a disposable demonstration sandbox, proves the
+healthy check can pass, then proves the hollow check is rejected. It writes a
+receipt and a privacy-safe Proof Card; it does not assess or change your project.
+
+### Pick the path that matches your work
+
+- **Individual developer or vibe coder:** start with `factory first-proof`.
+  Learn what a real negative control looks like before trusting AI-generated tests.
+- **Engineering team:** put `factory wrap` around an admitted Codex, Claude Code,
+  DeepSeek, or other local agent run. Code Factory records the exact file delta,
+  runs the validators your team declared, and stores hashes and bounded facts—not
+  prompts or raw model output—so reviewers can compare the claim with the change.
+- **Platform or assurance team:** evaluate local policy gates, evidence packets,
+  expiring exceptions, and tenant boundaries in a controlled pilot. Code Factory
+  is beta software; it does not provide a support SLA, compliance certification,
+  customer references, or procurement readiness.
+
+When you need more than the first proof, open the [advanced workflow guide](docs/OVERVIEW.md).
+To create a contained starter later, run
+`factory mvp "Build an approval tracker" --root .`.
+That starter is never called production-ready until the relevant proof exists.
 
 ![Actual privacy-safe First Proof Card: the hollow test was detected](docs/assets/first-proof-card.svg)
 
@@ -41,7 +57,7 @@ are ready to build, run `factory mvp "Build an approval tracker" --root .`.
 
 ![Actual Factory Studio showing the outcome-first local MVP path and local-only boundary](docs/assets/marketplace/factory-studio-mvp-1280x800.png)
 
-## What it does
+## Advanced workflows
 
 - **Start a real project.** Turn one outcome into a contained web, mobile, API,
   CLI, worker, MCP, or agent-UI starting state.
@@ -77,6 +93,13 @@ are ready to build, run `factory mvp "Build an approval tracker" --root .`.
   review date, and hash-bound proof obligations. A Change Safety Case routes
   one explicit diff to its owner; it never infers intent, promotes a decision,
   runs a test, or approves a release.
+- **Prove the whole journey—and audit the repairing agent.** Use
+  [`factory journey`](docs/JOURNEY_PROOF_ENGINE.md) to compare declared and
+  observed product journeys, preserve a hash-bound failure capsule, prove
+  state flow and cleanup, and challenge a proposed healing with a negative
+  mutation. Choose human-controlled or bounded supervised-auto mode; every
+  autonomous attempt receives a separate FactoryLine audit and neither mode
+  can self-approve.
 
 ## What the latest upgrade resolves
 
@@ -88,6 +111,7 @@ are ready to build, run `factory mvp "Build an approval tracker" --root .`.
 | Teams gradually trust an agent because it has been successful recently | Earned Autonomy derives an expiring local tier from governed evidence and demotes severe hollow-test or scope-escape results | Identity, permissions, execution, approval, and release authority |
 | A hard-won design decision is forgotten, then a later diff silently breaks its assumptions | Engineering Judgment Capsules bind one named owner, explicit path scope, review date, and declared proof obligations; an optional human-declared Change Profile makes novel boundaries and required Senior Attention explicit before review | Proposal, independent promotion, decision reconsideration, proof execution, acceptance, merge, and release |
 | A reviewer has suggestions but no shared evidence picture | Graph Ops, local receipts, and read-only MCP facts show current scope, proof debt, and the next fact-derived action | Merge, release, deployment, and provider access |
+| A selector repair makes the test green but changes the user journey—or an autonomous agent silently edits outside scope | Journey Reality, Stateful Workflow Proof, and Proof-Gated Healing require preserved semantic identity and coverage, cleanup proof, an adversarial negative result, exact workspace deltas, and an independent Agent Work Audit | Agent/provider choice, repair admission, final approval, merge, and release |
 | Agent work happens outside the evidence ledger, while writing real E2E manifests takes too long | Evidence Supply Line wraps any admitted local agent CLI; `gauntlet draft` proposes inert, structure-derived promise drafts and explicitly withholds commands it cannot derive | Agent identity, sandboxing, draft promotion, validator choice, Gauntlet admission, and release |
 | A large/remote workspace feels opaque or sluggish | Workspace Advisor measures bounded local project shape and path-only WSL/remote signals without changing indexes, heap, caches, inspections, or remote settings | Every IDE performance change and environment setting |
 
@@ -98,11 +122,11 @@ security, productivity, production readiness, or an automatic repair service.
 
 | Who is using it | Start here | Highest-value uses | The result they can inspect |
 | --- | --- | --- | --- |
-| **Novice / first-time builder** | `factory first-proof`, then `factory mvp` | Catch a test that only looks green; turn one outcome into a contained starter; learn the difference between generated code and proven behavior | A privacy-safe Proof Card, a local MVP, and a visible next step instead of an unexplained pass/fail |
+| **Individual developer / vibe coder** | `factory first-proof` | See whether a passing test can actually reject a known failure before trusting it | A local receipt and privacy-safe Proof Card from a disposable demonstration; your project remains unchanged |
 | **Junior developer** | `factory prd grill`, `factory plan verify`, `factory change review` | Clarify acceptance evidence before coding; keep an AI-assisted diff inside the approved plan; surface missing tests and Proof Debt before review | Source-bound questions, exact changed paths, severity-ordered findings, and a review handoff |
 | **Senior / staff engineer** | `factory judgment`, `factory graph forensics`, `factory proofsearch`, `factory gauntlet` | Protect architecture decisions; diagnose resumed or parallel workflow drift; compare candidate repairs; challenge whether critical E2E checks can actually reject declared failures | Hash-bound decision context, first-divergence facts, rejected candidates, a deterministic winner, and Survival Cards |
-| **Engineering team** | `factory wrap`, GitHub Proof Review, Graph Ops, Agent License and Combine | Put agent work into one evidence ledger; add a neutral PR Check beside AI review; hand off proof state; compare sealed runs without turning scores into vendor claims | Commit-bound Checks, local receipts, explicit owners and blockers, proof reuse decisions, and comparable governed-run records |
-| **Enterprise / platform team** | Intake Grill, Verifier Plane, policy challenge, assurance dossier, admission controls | Standardize approval boundaries; test whether policy rules matter; keep signers, exceptions, tenants, budgets, and release evidence explicit; integrate with existing SDLC controls | Independently verifiable packets and read-only operational views; identity, credentials, merge, release, and deployment remain in enterprise-owned systems |
+| **Engineering team** | `factory wrap`, then GitHub Proof Review | Record the exact agent-created file delta, run declared independent validators, and hand reviewers a commit-bound evidence packet without retaining prompts | Local receipts, exact changed paths, explicit blockers, and a neutral PR Check beside AI review |
+| **Platform / assurance pilot** | Policy gates and evidence packets | Evaluate named approval boundaries, expiring exceptions, tenant isolation, budgets, and release evidence inside existing SDLC controls | Independently verifiable packets and read-only views; identity, credentials, merge, release, and deployment remain enterprise-owned. No SLA, certification, references, or procurement-readiness claim |
 
 ## Expected operational outcomes
 
@@ -337,7 +361,7 @@ context reusable by a client you choose.
   optional design-quality lane and its explicit review boundaries.
 - Read [The approval signal decays when AI-written code becomes routine](docs/HABITUATION_ESSAY.md)
   for the design and limits of the habituation gate.
-- See the [release notes](docs/RELEASE_NOTES_0.44.2.md),
+- See the [release notes](docs/RELEASE_NOTES_0.44.3.md),
   [CHANGELOG.md](CHANGELOG.md), [release channels](docs/RELEASE_CHANNELS.md), and
   [publication guide](PUBLICATION_GUIDE.md) for versioned release detail.
 

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/AI_CLIENTS.md

@@ -11,7 +11,7 @@ Install the published package in the same environment that the client can
 launch, then verify the workspace boundary:
 
 ```powershell
-python -m pip install factoryline-code-factory==0.44.2
+python -m pip install factoryline-code-factory==0.44.3
 factory mcp status --root C:\work\my-mvp --json
 factory mcp config --client generic --root C:\work\my-mvp --json
 ```

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/FIRST_USE.md

@@ -5,7 +5,7 @@ small feature or risky workflow, not a whole migration.
 
 ```powershell
 cd path\to\your-repository
-pip install factoryline-code-factory==0.44.2 code-factory-1-spec==0.5.4 code-factory-2-forge==0.10.7
+pip install factoryline-code-factory==0.44.3 code-factory-1-spec==0.5.4 code-factory-2-forge==0.10.7
 factory doctor --json
 factory init .
 forge adopt <feature> --root .

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/INTELLIJ.md

@@ -133,7 +133,7 @@ candidate files.
 1. Install FactoryLine:
 
    ```powershell
-   pip install factoryline-code-factory==0.44.2
+   pip install factoryline-code-factory==0.44.3
    ```
 
 2. In your JetBrains IDE, open **Settings > Plugins > Marketplace**, search for

factoryline_code_factory-0.44.3/docs/JOURNEY_PROOF_ENGINE.md

@@ -0,0 +1,97 @@
+# Journey Proof Engine
+
+Code Factory's Journey Proof Engine checks whether the product journey that was
+declared is the journey that the runtime actually exposed. It is local,
+provider-neutral, and deterministic: browsers, APIs, CI systems, and external
+test runners may supply observations, but none receives approval or release
+authority.
+
+## What it adds
+
+### Journey Reality Graph
+
+`factory journey reality` compares declared and observed states, transitions,
+requirements, outcomes, and SHA-256-bound runtime artifacts. It reports exact
+missing, unexpected, and stale items. It does not ask a model to guess whether
+two differently named states mean the same thing.
+
+### Rich Failure Capsules
+
+`factory journey capsule` preserves the failed step, at most one adjacent step
+on each side, a closed failure classification, workspace-contained artifacts,
+and separate unverified hypothesis and repair fields. It writes a JSON receipt
+and a reviewer-friendly Markdown capsule without claiming a root cause.
+
+### Stateful Workflow Proof
+
+`factory journey workflow-proof` verifies the complete workflow rather than
+isolated green tests. Every consumed value needs one earlier producer with the
+same hash. Every created side effect needs a later successful cleanup, and each
+cleanup needs a passing idempotency probe.
+
+### Proof-Gated Healing
+
+`factory journey heal-verify` admits a candidate repair for review only when:
+
+- the patch and changed paths are SHA-256-bound and allowlisted;
+- role, label, route, and state anchors retain semantic identity;
+- previously covered journeys remain covered;
+- the positive command exits zero; and
+- the adversarial negative mutation exits non-zero.
+
+A negative command that also passes is a hollow healing proof and fails closed.
+
+## Human and autonomous modes
+
+The same proof engine supports two explicit modes:
+
+- `human_controlled` forbids an agent command. It verifies the prepared repair
+  and returns `HEALING_HUMAN_REVIEW_REQUIRED`.
+- `supervised_auto` permits one to three no-shell attempts by an explicitly
+  identified local command. Code Factory snapshots the workspace before and
+  after every attempt, stops on the first scope escape, and returns
+  `HEALING_AUTO_AWAITING_PROMOTION` only after the proof succeeds.
+
+Neither mode grants final approval. Graph Ops can create an inert manifest
+template, but a separately reviewed terminal must run it.
+
+## The agent is audited too
+
+Every supervised attempt produces a separate
+`factory.agent-work-audit.v1` receipt. FactoryLine—not the worker—binds:
+
+- the declared agent identity and argv digest;
+- before and after workspace digests;
+- exact changed paths and scope verdict;
+- the agent, positive, and negative-mutation results; and
+- one closed audit outcome and, for a failed outcome, one classification from
+  Code Factory's closed `FailureClass` taxonomy.
+
+The receipt stores hashes of bounded stdout and stderr rather than raw,
+potentially sensitive logs. Worker approval is always `false`.
+
+## BYOK and managed execution
+
+BYOK/local execution is the default OSS path: private, provider-neutral, and
+without hidden token markup. A managed service can later supply hosted keys,
+usage controls, audit retention, team policy, support, and predictable billing.
+Both paths must produce the same inputs and pass the same independent proof
+engine; billing or provider choice cannot weaken the verifier.
+
+## Read-only inspection
+
+```bash
+factory journey status --root . --json
+```
+
+MCP clients can call `factory.journey_status`. Graph Ops projects only
+hash-verified receipts and links Agent Work Audits to their healing decisions.
+Both surfaces expose `JOURNEY_STATUS_READ_ONLY` and cannot run agents, apply
+repairs, approve, merge, publish, deploy, sign, message, access credentials, or
+grant connectors.
+
+## Receipt location
+
+Receipts are written atomically under `.factory/journey-proof/`. Each receipt
+contains a canonical `receipt_sha256`, an all-false authority map, and the exact
+decision facts needed for offline review.

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/LANGGRAPH_ASSURANCE.md

@@ -89,7 +89,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v5
-      - uses: zrk222/code-factory@v0.44.2
+      - uses: zrk222/code-factory@v0.44.3
         with:
           reference: .factory/langgraph/reference.json
           resumed: .factory/langgraph/resumed.json

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/MCP_REGISTRY.md

@@ -2,7 +2,7 @@
 
 Code Factory is published to the Official MCP Registry as
 `io.github.zrk222/code-factory`. The registry package entry uses the public
-PyPI distribution `factoryline-code-factory==0.44.2` and starts the existing
+PyPI distribution `factoryline-code-factory==0.44.3` and starts the existing
 local stdio adapter through `uvx`:
 
 **Use it to give an AI coding client read-only facts about declared intent,
@@ -10,7 +10,7 @@ test challenges, Graph Ops, and review evidence—without handing that client
 write or release authority.**
 
 ```text
-uvx --from factoryline-code-factory==0.44.2 factory mcp serve
+uvx --from factoryline-code-factory==0.44.3 factory mcp serve
 ```
 
 The server needs a workspace root. Configure that explicit path in a client

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/OPENVSX.md

@@ -26,7 +26,7 @@ verified before that provider action occurs.
    repository, then store a least-privilege `OPENVSX_TOKEN` in the protected
    `openvsx` GitHub environment.
 2. Start **Publish Open VSX extension** manually and supply an existing,
-   immutable repository tag such as `v0.44.2`.
+   immutable repository tag such as `v0.44.3`.
 3. Leave `publish` false to produce and retain only the verified VSIX candidate,
    SHA-256 manifest, audit result, and tests.
 4. Set `publish` true only after the environment reviewer approves the target

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/OVERVIEW.md

@@ -89,7 +89,7 @@ known, what is missing, and the one fact-derived next action.
 ## Typical path
 
 ```powershell
-pip install factoryline-code-factory==0.44.2
+pip install factoryline-code-factory==0.44.3
 factory mvp "Build an approval tracker" --root .
 factory studio --root .
 factory graph ops --root . --mermaid

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/PRODUCT_HUNT_GALLERY.md

@@ -45,36 +45,36 @@ gallery replacement.
 Use this only as an update associated with the existing **Code Factory**
 Product Hunt product.
 
-**Headline:** Refresh the proof state. See the next safe action.
+**Headline:** Can your AI-generated test actually fail?
 
-**Tagline:** Follow live local delivery evidence without handing an AI release authority.
+**Tagline:** One local command exposes a test that only looks green.
 
 **Update copy:**
 
-Code Factory now keeps Factory Studio and Unified Graph Ops current while a
-local Assembly runs. Refresh the activity state, see the current stage and
-completed work, and open a Developer Memory Brief that turns the exact diff into
-one safe next-proof action. It also shows redacted continuity facts and observed
-local Git contribution context—never a claim that Git authors are verified
-seats, owners, approvers, or a billing roster.
+I built Code Factory after losing time to AI-generated tests that passed but
+failed in real use. The front door is now deliberately simple: run
+`factory first-proof --root .` and see a safe demonstration of whether a test
+can actually reject a known failure. It stays local and does not change your
+project.
 
-Evidence Frontier then ranks the supplied, non-executing checks by how many
-viable repair pairs each can distinguish. Its predictions are displayed as
-unverified hypotheses—not a model verdict—and the next experiment remains
-locked until a separately approved runner records external evidence.
+For a solo developer or vibe coder, that is the whole starting value. For a
+team, `factory wrap` records the exact file delta from an admitted agent run,
+runs the validators the team declared, and leaves a reviewable receipt without
+retaining prompts or raw model output. Platform and assurance teams can evaluate
+the deeper policy and evidence controls in a pilot, but this beta does not claim
+an SLA, certification, customer references, or procurement readiness.
 
-The same UI still makes evidence and authority visible: fresh or stale proof,
-next action, and measured-or-unavailable values are separated from approvals.
-Verification, export, and guardrail controls remain available; tests, repairs,
-apply, merge, publication, deployment, credentials, and connectors remain
-human-controlled.
+Version 0.44.3 also compares declared and observed product journeys, preserves
+the failed step in a bounded evidence capsule, checks state and cleanup, and
+audits the repairing agent separately. Human approval, apply, merge, release,
+credentials, and deployment remain outside Code Factory's authority.
 
-**Primary CTA:** Inspect the live repair arena and verify the winning evidence.
+**Primary CTA:** Run your first local proof.
 
 **Proof links:**
 
-- GitHub release: `https://github.com/zrk222/code-factory/releases/tag/v0.44.2`
-- PyPI: `https://pypi.org/project/factoryline-code-factory/0.44.2/`
+- GitHub release: `https://github.com/zrk222/code-factory/releases/tag/v0.44.3`
+- PyPI: `https://pypi.org/project/factoryline-code-factory/0.44.3/`
 - Source: `https://github.com/zrk222/code-factory`
 
 If no authenticated Code Factory post or draft exists, report

{factoryline_code_factory-0.44.2 → factoryline_code_factory-0.44.3}/docs/RELEASE_CHANNELS.md

@@ -1,13 +1,13 @@
 # Release Channels
 
-Code Factory v0.44.2 publishes one verified source state through channel-specific
+Code Factory v0.44.3 publishes one verified source state through channel-specific
 adapters. A successful GitHub release is not evidence that every moderated
 listing has accepted the artifact.
 
 | Channel | Artifact or surface | Release path | Success evidence |
 | --- | --- | --- | --- |
-| GitHub | Source tag, wheel, sdist, VSIX, JetBrains ZIP, media | Publish `v0.44.2`; `publish.yml` attaches the verified bundle | Public release URL and green workflow |
-| PyPI | `factoryline-code-factory==0.44.2` | Trusted Publishing from `publish.yml` | PyPI project version and attestation |
+| GitHub | Source tag, wheel, sdist, VSIX, JetBrains ZIP, media | Publish `v0.44.3`; `publish.yml` attaches the verified bundle | Public release URL and green workflow |
+| PyPI | `factoryline-code-factory==0.44.3` | Trusted Publishing from `publish.yml` | PyPI project version and attestation |
 | Official MCP Registry | `io.github.zrk222/code-factory` local stdio descriptor | Post-PyPI GitHub OIDC job in `publish.yml` | Public registry entry and green registry job |
 | Hugging Face | Static Code Factory Space | Push `deploy/huggingface/` to `main` | Green Space workflow and public Space |
 | Zenodo | Versioned source archive under concept DOI | GitHub release integration | Public version record; concept DOI remains stable |
@@ -19,6 +19,16 @@ The release pipeline never treats a queued review, draft listing, uploaded
 artifact, or workflow dispatch as a completed publication. Each channel is
 reported as published, pending review, blocked, or not configured.
 
+### 0.44.3 Journey proof and audited healing
+
+This release adds Journey Reality, bounded failure capsules, and stateful
+workflow proof. Proof-gated healing can remain human-controlled or run in a
+supervised automatic mode, but a candidate is accepted only when its positive
+proof passes, its negative control fails, and the repairing agent's work passes
+an independent audit. Unknown evidence stays unknown. BYOK/local remains the
+default, and neither mode grants approval, merge, release, or credential
+authority.
+
 ### 0.44.2 Intent lineage navigation
 
 This patch makes a verified Forge intent trace easier to inspect in Graph Ops.