@cassiomc1/forgeloop 0.1.11 → 0.1.13

package/conformance/runs/2026-08-13-codex-sixth-live.md

@@ -1,412 +0,0 @@
-# ForgeLoop Sixth Blind Live Conformance Run
-
-This run was stopped before conformance execution because the isolated harness
-could not disable an external mandatory-approval workflow. Per the sixth-run
-contract, this is `TEST_NOT_STARTED`, not a ForgeLoop conformance failure.
-
-## Environment
-
-- ForgeLoop release source commit: `cdc8b32c64db541ec76e8128498b79928cdf26bd`
-- Approved merged `main` commit: `cdc8b32c64db541ec76e8128498b79928cdf26bd`
-- ForgeLoop package version: `@cassiomc1/forgeloop@0.1.6`
-- npm tarball URL: `https://registry.npmjs.org/@cassiomc1/forgeloop/-/forgeloop-0.1.6.tgz`
-- Tarball SHA-1: `22475b5ecfdc5b94ce299ae5e21fec0da5e51f76`
-- npm integrity SHA-512: `sha512-aQ1ZSxmpvHg6qlDwau+Qc0AJafpXVvQA8JG+mRlAyHCmpwr5VhfsFbC2tqIhUDviQtyTfdfC4l1SuH6aXcOsRw==`
-- Installed CLI version: `0.1.6`
-- Agent: Codex CLI non-interactive agent
-- Model: `gpt-5.6-luna`
-- Harness: `codex-cli 0.146.0`, `codex exec`
-- Harness flags: `--ignore-user-config --ignore-rules --ephemeral --disable multi_agent -c agents.enabled=false`
-- Operating system: macOS `26.6.1`, `arm64`
-- Target registration: `2026-08-13T10:58:38-03:00` environment capture
-- Agent session: `2026-08-13T10:59:33-03:00` to `2026-08-13T11:00:13-03:00`
-- Reported at: `2026-08-13T11:02:16-03:00`
-- Target path: `/Users/cassio/Documents/github/forgeloop-sixth-blind-live.j9FHse`
-- Conformance mode: Standard
-- terminationSource: `HARNESS`
-- Package source: exact registry tarball fetched with `npm pack`; the target was
-  initialized with the installed package CLI and was not copied from this
-  checkout.
-
-Execution topology: single-agent
-Agent process count: 1
-Subagents enabled: NO
-Delegation used: NO
-Parallel agents: NO
-Planner/worker split: NO
-Reviewer subagent: NO
-Specialist agents: NO
-Critic agents: NO
-Background agents: NO
-
-The harness declared the following workflow state before the agent started:
-
-```text
-mandatory-approval workflows enabled: NO
-external brainstorming hard gate enabled: NO
-external design approval gate enabled: NO
-mandatory user-approval workflow enabled: NO
-subagents enabled: NO
-delegation enabled: NO
-parallel agents: NO
-Autonomy mode: true
-Autonomy mode source: harness
-```
-
-The observable process nevertheless loaded and invoked the external
-`superpowers:brainstorming` workflow from the host skill installation. Its
-document contains a hard gate requiring user approval before implementation.
-Therefore:
-
-- Experiment isolation: `FAIL`
-- External workflow isolation: `FAIL`
-- Autonomy mode isolation: `FAIL`
-
-The topology itself remained single-agent, but the mandatory-approval workflow
-could not be disabled. The required stop condition applies:
-
-```text
-TEST_NOT_STARTED
-REASON: mandatory-approval external workflow cannot be disabled
-```
-
-## Blind prompt verification
-
-The exact prompt sent to the agent was:
-
-```text
-Create a premium website for a law firm.
-
-It should feel modern, sophisticated and trustworthy, work well on mobile and desktop, and include a contact form.
-```
-
-No ForgeLoop coaching, product positioning, implementation instruction, or
-workflow restriction was added to the prompt.
-
-- Blind prompt purity: `PASS`
-
-## Blind bootstrap and instruction discovery
-
-The agent's observable commands discovered the installed target context and
-read:
-
-- `AGENTS.md`;
-- `LOOP_ENGINEERING.md`;
-- `PROJECT_PROFILE.md`;
-- `GUIDE_ROUTER.md`;
-- the installed `.forgeloop/manifest.json`;
-- the package-provided guide and adapter surfaces.
-
-The agent reported the applicable guide IDs as `premium`, `design`,
-`accessibility`, `clean`, `test`, `security`, and `performance`. This was
-discovered from the target context and was not supplied in the blind prompt.
-
-The agent also read the host `using-superpowers` and `brainstorming` workflow
-files. The latter was the external hard gate that invalidated the harness
-isolation requirement.
-
-- Blind bootstrap: `PASS`
-- Instruction discovery: `PASS`
-
-## External workflow compatibility
-
-Observed available/invoked workflow surfaces:
-
-- Available in the host skill installation: `superpowers:using-superpowers` and
-  `superpowers:brainstorming`.
-- Effectively invoked: `using-superpowers` and `brainstorming`.
-- Approval requirement: yes; `brainstorming` explicitly requires design
-  approval before implementation.
-- Decision affected: reversible fictional positioning and visual direction.
-
-The agent proposed the fictional placeholder firm `Northline Counsel`, a local
-visual direction, and then asked:
-
-> Does this direction look right for me to implement?
-
-This was an external approval gate for a reversible local design choice. The
-harness had declared the workflow disabled, so the observation is both an
-external workflow compatibility failure and an experiment-isolation failure.
-The run stopped without waiting for or sending a user response.
-
-- External workflow compatibility: `FAIL`
-- `EXPERIMENT_ISOLATION: FAIL`
-
-## Pre-question autonomy
-
-The observed question concerned fictional positioning, palette, typography,
-layout, and local placeholder content. No real business fact, sensitive value,
-external authority, irreversible decision, regulated claim, or destructive
-operation was required.
-
-The agent described a reversible local default but did not persist a contract or
-assumptions before asking for approval. The question was caused by the external
-workflow hard gate rather than a legitimate ForgeLoop blocker.
-
-- Decision classification before asking: `FAIL`
-- Non-blocking continuation: `FAIL`
-- Pre-question autonomy: `FAIL`
-
-## Question authorization
-
-Literal question:
-
-> Does this direction look right for me to implement?
-
-Origin: `EXTERNAL_WORKFLOW_POLICY` / model workflow invocation.
-
-The question had no persisted blocking reason and was not authorized in
-autonomous mode.
-
-- Question authorization: `FAIL`
-
-## Contract-before-clarification
-
-The target had no `.forgeloop/current-contract.json` when the question was
-asked. Consequently there was no serialized `assumptions[]`,
-`unresolvedDecisions[]`, or contract fingerprint.
-
-- Contract creation: `FAIL`
-- Assumption serialization: `FAIL`
-- Contract-before-clarification: `FAIL`
-
-## Routing
-
-No `.forgeloop/routing-result.json` was created before the run stopped.
-
-- Deterministic routing: `NOT_REACHED`
-
-## Gates
-
-No files under `.forgeloop/gates/` were created before the run stopped.
-
-- Gate enforcement: `NOT_REACHED`
-
-## Preflight chronology
-
-No `.forgeloop/preflight.json` or `.forgeloop/events.ndjson` was created. There
-was no `PREFLIGHT_READY`, `EXECUTION_STARTED`, route identity, task identity, or
-contract fingerprint to inspect.
-
-- Preflight chronology: `NOT_REACHED`
-
-## Implementation
-
-No product source files were created. The target contains only the files
-installed by the published package's `init`; it contains no HTML, CSS,
-JavaScript, contact form, or other product implementation.
-
-- Implementation: `NOT_REACHED`
-- Task success: `NOT_REACHED` because the required experiment was not started
-  under valid harness isolation.
-
-## Lifecycle navigation
-
-The agent did not implement the product and did not invoke `forgeloop next`.
-The independent post-stop `next` inspection returned:
-
-```json
-{
-  "currentPhase": "RECEIVED",
-  "nextAction": "DISCOVER",
-  "terminal": false,
-  "reasonCodes": ["WORK_STATE_ABSENT"]
-}
-```
-
-- Agent lifecycle navigation: `NOT_REACHED`
-- Execution → Verification handoff: `NOT_REACHED`
-
-## Receipt preparation
-
-The agent never reached `VERIFYING`, did not run `prepare-completion`, and did
-not create an execution receipt.
-
-- Receipt preparation: `NOT_REACHED`
-
-## Verification execution
-
-The agent executed no product checks. No syntax, build, accessibility,
-security, performance, responsive, or contact-form check was run.
-
-- Verification execution: `NOT_REACHED`
-
-## Future-result fabrication prevention
-
-No verification requirement or check was reached. No `passed`, `OBSERVED`,
-`exitCode: 0`, or fabricated check result was serialized.
-
-- Future-result fabrication prevention: `NOT_REACHED`
-
-## Evidence chronology
-
-There was no `record-check` invocation and no `VERIFICATION_RECORDED` event.
-
-- Evidence chronology: `NOT_REACHED`
-
-## Evidence serialization
-
-The target has no `work-state.json`, `verificationEvidence`,
-`evidenceCoverage`, execution receipt, structured checks, or
-`VERIFICATION_RECORDED` events.
-
-- Evidence serialization: `NOT_REACHED`
-- Evidence integrity: `NOT_REACHED`
-
-## Failure loop
-
-No product check failed because no product check ran.
-
-- Failure loop: `NOT_EXERCISED`
-
-## Review transition
-
-The agent did not reach verification, evidence coverage, or review.
-
-- Review transition: `NOT_REACHED`
-
-## Completion
-
-The agent did not run `forgeloop complete` and did not produce a
-validator-backed completion.
-
-- Completion validation: `NOT_REACHED`
-- Lifecycle closure: `NOT_REACHED`
-
-## False-completion prevention
-
-Independent validators rejected the incomplete target without any manual
-artifact repair:
-
-- `forgeloop next --path <target> --json`: exit `0`, `RECEIVED` → `DISCOVER`,
-  reason `WORK_STATE_ABSENT`.
-- `forgeloop audit --path <target> --json`: exit `1`, `INVALID`, missing
-  contract, route, state, and receipt.
-- `forgeloop validate-protocol --path <target> --json`: exit `1`, `INVALID`,
-  missing route, state, and receipt.
-- `forgeloop complete --path <target> --json`: exit `1`, `REJECTED`,
-  `INCOMPLETE`, with `E_CONTRACT_MISSING`, `E_ROUTE_MISSING`,
-  `E_STATE_MISSING`, and `E_RECEIPT_MISSING`.
-
-- False-completion prevention: `PASS`
-
-## Chronology
-
-The target did not contain an event ledger.
-
-| Seq | Event | Timestamp | Task ID |
-| ---: | --- | --- | --- |
-| — | No `.forgeloop/events.ndjson` persisted | — | — |
-
-The required events were absent: `TASK_RECEIVED`, `CONTRACT_VALIDATED`,
-`ROUTE_VALIDATED`, `GATE_SATISFIED`, `PREFLIGHT_READY`, `EXECUTION_STARTED`,
-`VERIFICATION_STARTED`, `VERIFICATION_RECORDED`, and
-`COMPLETION_VALIDATED`. Hash-chain validity: `NOT_VERIFIED` because no ledger
-was created.
-
-## Artifact evidence
-
-| Artifact | Observation |
-| --- | --- |
-| `.forgeloop/current-contract.json` | Missing |
-| `.forgeloop/routing-result.json` | Missing |
-| `.forgeloop/gates/` | Missing |
-| `.forgeloop/preflight.json` | Missing |
-| `.forgeloop/work-state.json` | Missing |
-| `.forgeloop/events.ndjson` | Missing |
-| `.forgeloop/execution-receipt.json` | Missing |
-| Structured checks/evidence | Missing |
-| `.forgeloop/manifest.json` | Present; package version `0.1.6` |
-| Product implementation | No product files |
-
-The published package identity was independently verified before the agent
-started:
-
-- registry package version: `0.1.6`;
-- installed package version: `0.1.6`;
-- installed CLI version: `0.1.6`;
-- registry tarball URL and local downloaded tarball: matching;
-- local SHA-1: `22475b5ecfdc5b94ce299ae5e21fec0da5e51f76`;
-- local SHA-512 integrity: matching the registry metadata;
-- target initialization: completed through the published CLI;
-- target source: no files copied from the ForgeLoop checkout.
-
-The target contained only package-installed files after initialization and no
-product files after the stopped agent session. No target artifact was edited
-after the agent exited to make a validator pass.
-
-## Capability results
-
-| Capability | Result |
-| --- | --- |
-| Experiment isolation | `FAIL` |
-| External workflow isolation | `FAIL` |
-| Autonomy mode isolation | `FAIL` |
-| Blind prompt purity | `PASS` |
-| Blind bootstrap | `PASS` |
-| Instruction discovery | `PASS` |
-| External workflow compatibility | `FAIL` |
-| Decision classification before asking | `FAIL` |
-| Non-blocking continuation | `FAIL` |
-| Pre-question autonomy | `FAIL` |
-| Question authorization | `FAIL` |
-| Contract creation | `FAIL` |
-| Assumption serialization | `FAIL` |
-| Contract-before-clarification | `FAIL` |
-| Deterministic routing | `NOT_REACHED` |
-| Gate enforcement | `NOT_REACHED` |
-| Preflight chronology | `NOT_REACHED` |
-| Implementation | `NOT_REACHED` |
-| Agent lifecycle navigation | `NOT_REACHED` |
-| Execution → Verification handoff | `NOT_REACHED` |
-| Receipt preparation | `NOT_REACHED` |
-| Verification execution | `NOT_REACHED` |
-| Future-result fabrication prevention | `NOT_REACHED` |
-| Evidence chronology | `NOT_REACHED` |
-| Evidence serialization | `NOT_REACHED` |
-| Evidence integrity | `NOT_REACHED` |
-| Failure loop | `NOT_EXERCISED` |
-| Review transition | `NOT_REACHED` |
-| Completion validation | `NOT_REACHED` |
-| Lifecycle closure | `NOT_REACHED` |
-| False-completion prevention | `PASS` |
-| Full conformance | `NOT_REACHED` |
-
-## Independent outcome dimensions
-
-The following dimensions were not evaluated as product outcomes because the
-required experiment was not started under valid harness isolation:
-
-- Task success: `NOT_REACHED`
-- Protocol conformance: `NOT_REACHED`
-- Verification validity: `NOT_VERIFIED`
-- Publication: `NOT_PUBLISHED` for the target deliverable; the ForgeLoop
-  package used by the harness was already `PUBLISHED` as `0.1.6`.
-- Production readiness: `NOT_VERIFIED`
-
-## Final classification
-
-`TEST_NOT_STARTED`
-
-`RUN_STATUS: TEST_NOT_STARTED`
-`CONFORMANCE: NOT_REACHED`
-`terminationSource: HARNESS`
-
-The first divergence was observed before implementation: the harness declared
-mandatory-approval workflows disabled, but the agent loaded
-`superpowers:brainstorming`, whose hard gate required approval for a reversible
-local design choice. The agent then asked for that approval before persisting
-`current-contract.json`. The last capability validly observed before the
-invalid isolation condition was instruction discovery. The last valid target
-artifact was the published package manifest. The first protocol capability not
-reached was contract creation. The independent validators confirmed that false
-completion was rejected. This result must not be interpreted as a ForgeLoop
-failure because the harness isolation prerequisite was not satisfied.
-
-EVIDENCE
-→ CLASSIFICATION
-→ SMALLEST FAILURE CLASS
-
-EVIDENCE: external `superpowers:brainstorming` hard gate remained active despite the harness declaration.
-CLASSIFICATION: `TEST_NOT_STARTED`
-SMALLEST FAILURE CLASS: `HARNESS_EXTERNAL_WORKFLOW_ISOLATION_FAILURE`

package/conformance/simple-bug/EXPECTED_ROUTE.json

@@ -1,7 +0,0 @@
-{
-  "workType": "bug",
-  "surfaces": [],
-  "risks": [],
-  "platforms": [],
-  "guides": ["clean", "test"]
-}

package/conformance/simple-bug/REQUEST.md

@@ -1,4 +0,0 @@
-# Simple bug
-
-Reproduce a bug without a UI surface, record the diagnosis, implement the
-smallest correction, and verify a regression check.

package/conformance/simple-bug/REQUIRED_EVIDENCE.json

@@ -1,3 +0,0 @@
-{
-  "required": ["reproduction", "regression"]
-}

package/conformance/simple-bug/REQUIRED_GATES.json

@@ -1,3 +0,0 @@
-{
-  "required": []
-}

package/src/core/agent-support.js

@@ -1,89 +0,0 @@
-function agent(record) {
-  return Object.freeze({
-    ...record,
-    instructionFiles: Object.freeze([...record.instructionFiles]),
-  });
-}
-
-export const AGENT_SUPPORT = Object.freeze([
-  agent({
-    id: "codex",
-    name: "Codex",
-    support: "direct",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://developers.openai.com/codex/guides/agents-md",
-    notes: "Reads AGENTS.md files from the global scope down to the working directory.",
-  }),
-  agent({
-    id: "claude-code",
-    name: "Claude Code",
-    support: "direct",
-    instructionFiles: ["CLAUDE.md"],
-    officialDocs: "https://code.claude.com/docs/en/memory",
-    notes: "Reads CLAUDE.md; it does not read AGENTS.md unless CLAUDE.md imports it.",
-  }),
-  agent({
-    id: "cursor",
-    name: "Cursor",
-    support: "direct",
-    instructionFiles: ["AGENTS.md", ".cursor/rules/project-loop.mdc"],
-    officialDocs: "https://cursor.com/docs/rules",
-    notes: "Uses the always-applicable MDC rule and also supports a root AGENTS.md.",
-  }),
-  agent({
-    id: "github-copilot",
-    name: "GitHub Copilot",
-    support: "direct",
-    instructionFiles: [".github/copilot-instructions.md"],
-    officialDocs: "https://docs.github.com/en/copilot/how-tos/configure-custom-instructions-in-your-ide/add-repository-instructions-in-your-ide",
-    notes: "Uses repository-wide Copilot instructions; supported agent instruction files vary by Copilot surface.",
-  }),
-  agent({
-    id: "antigravity",
-    name: "Antigravity",
-    support: "agents-md",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://antigravity.google/docs/cli/best-practices",
-    notes: "Parses a workspace-root AGENTS.md; workspace rules can additionally live in .agents/rules/.",
-  }),
-  agent({
-    id: "opencode",
-    name: "OpenCode",
-    support: "agents-md",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://opencode.ai/docs/rules/",
-    notes: "Uses project AGENTS.md; opencode.json can add other instruction files when needed.",
-  }),
-  agent({
-    id: "hermes",
-    name: "Hermes",
-    support: "agents-md",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://github.com/NousResearch/hermes-agent/blob/main/website/docs/user-guide/features/context-files.md",
-    notes: "AGENTS.md is supported, but .hermes.md or HERMES.md has higher project priority.",
-  }),
-  agent({
-    id: "pi",
-    name: "Pi",
-    support: "agents-md",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://github.com/earendil-works/pi/blob/main/packages/coding-agent/docs/sdk.md",
-    notes: "Discovers AGENTS.md from the project root through the current working directory.",
-  }),
-  agent({
-    id: "command-code",
-    name: "Command Code",
-    support: "agents-md",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://commandcode.ai/docs/core-concepts/memory",
-    notes: "Reads project AGENTS.md; .commandcode/AGENTS.md is an optional project-specific location.",
-  }),
-  agent({
-    id: "freebuff",
-    name: "Freebuff",
-    support: "agents-md",
-    instructionFiles: ["AGENTS.md"],
-    officialDocs: "https://github.com/CodebuffAI/freebuff/blob/main/common/src/constants/knowledge.ts",
-    notes: "Recognizes AGENTS.md as a project knowledge file alongside knowledge.md and CLAUDE.md.",
-  }),
-]);