# Workflow Compatibility Audit Reference workflow snapshot: `3fb8802e314dc0a1f23481dd3782cdca26b92dc2`. This document tracks the reference workflow surfaces checked during the deeper LitClaude compatibility pass. The goal is to keep future updates honest: if the reference grows a new user-facing skill, hook, or runtime behavior, this map should make the missing LitClaude target obvious. ## Workflow Parity Matrix | Reference workflow category | LitClaude target | Status | Reason / acceptance signal | | --- | --- | --- | --- | | review-work 5-lane orchestration | `commands/review-work.md`, `skills/review-work`, `agents/quality-reviewer.md`, `agents/lit-verifier.md`, `agents/qa-runner.md`, `agents/librarian-researcher.md` | Implemented | `review-work` represents scout / skeptic / implementer / edge-case / decision ledger discipline as scope/diff verification, tests/evidence execution, package/payload and code quality, security/provenance, and real-surface/docs readiness with an aggregate verdict. | | litgoal state/runtime | `commands/litgoal.md`, `skills/litgoal`, `plugins/litclaude/lib/litgoal/`, `.litclaude/litgoal/`, `bin/litclaude-ai.js litgoal ...` commands | Implemented | `litgoal` now has durable state, ledger, checkpoint, steering, quality gate, and review blocker commands through `litclaude litgoal create-goals`, `record-evidence`, `checkpoint`, `steer`, and `record-review-blockers`. | | LIT/start-work discipline | `commands/lit-loop.md`, `commands/lit-plan.md`, `commands/start-work.md`, `skills/lit-loop`, `skills/lit-plan`, `skills/start-work` | Implemented | Existing LIT/start-work surfaces require plan reading, ledger updates, publish boundary checks, verification, manual QA artifacts, cleanup receipts, resume checkpoints, and evidence artifacts. | | hook trigger safety | `hooks/hooks.json`, `bin/litclaude-hook.js`, `test/hooks.test.mjs`, `docs/hooks.md` | Implemented | Hooks use `UserPromptSubmit` trigger detection, exact workflow aliases, near-miss fixtures, prompt injection safe context, and no auto-type behavior for slash commands. | | bundled handoff | `skills/lit-handoff`, `commands/lit-handoff.md`, `vendor/handoff`, exact-bare hook | Implemented | Four canonical files remain byte-identical; exact bare `handoff` injects the complete source while mentions, compounds, slash text, and secret-bearing prompts remain inert. | | bundled scientific visualization | `skills/lit-scientific-visualization`, `commands/lit-scientific-visualization.md`, `vendor/scientific-visualization`, exact-bare hook, capability doctor | Implemented | Sixteen authored files remain byte-identical; the exact-bare hook displays the banner and injects the complete adapter/source while mixed and generic prompts stay inert; installed source-root path resolution, dependency DEGRADED status, Python source tests, and real figure export are covered without silent installation. | | canonical frontend library | `skills/frontend-ui-ux/references/{design,designpowers,perfection,ui-ux-db}`, `_canonical-corpus/manifest.json`, scanner, doctor, package guard | Implemented | Exact paths, sizes, hashes, legal companions, aggregate, entry types, no extras, scanner snapshot scope, and produced-tarball bytes are checked; imported content is inert and independent from the normalized dataset. | | autoresearch family | `skills/autoresearch`, `commands/autoresearch.md`, `vendor/autoresearch`, hook/doctor/package surfaces | Implemented | One Claude-native skill routes ten nested modes through plan, explicit budget/authority approval, start-work, bounded evidence loops, and review-work. Its reviewed commit is independently bound to a hardcoded tree commitment outside the generated closure manifest. | | autoconference family | `skills/autoconference`, `commands/autoconference.md`, `vendor/autoconference`, hook/doctor/package surfaces | Implemented | One skill routes seven modes, requires autoresearch and root multi-agent capability, keeps shared state root-owned, and blocks rather than simulating concurrency. Its reviewed commit is independently bound to a hardcoded tree commitment outside the generated closure manifest. | | wikify family | `skills/wikify`, `commands/wikify.md`, `vendor/llm-wikify`, hook/doctor/package surfaces | Implemented | One task-local skill routes five modes from litresearch through inert local sources to review-work and continuity. Its reviewed commit is independently bound to a hardcoded tree commitment outside the generated closure manifest. | | agent/command routing | `agents/*.md`, `commands/*.md`, `.claude-plugin/plugin.json`, command discovery tests | Implemented | Agents and commands route planner, executor, verifier, reviewer, librarian, and QA responsibilities into Claude Code surfaces, including review-work and litgoal command files. | | docs/package payload | `README.md`, `README_ko-KR.md`, `docs/*.md`, `package.json`, `.claude-plugin/plugin.json`, `test/workspace-hygiene.test.mjs` | Implemented | Docs, package payload, manifest, install metadata, package file-list, and guarded-token gates are covered by tests without implying publish completion. | | excluded/deferred telemetry | No runtime telemetry package; docs omission note | Omitted | Telemetry is omitted and any network telemetry remains deferred by reason: LitClaude should stay local and quiet unless the user explicitly requests an observable reporting feature. | The guarded-token scanner binds its verdict to one immutable in-memory snapshot of the enumerated inventory and exact bytes. The report carries that snapshot's file count and deterministic SHA-256 digest. Success does not prove the mutable live tree remained clean after capture; a later tree state requires another run, while package and tarball guards independently bind their own exact payloads. Capture is bounded to 8 MiB per file and 32 MiB in aggregate, and the zero-entry allowlist is parsed from the captured bytes rather than a pre-capture pathname read. Legal companion paths, source paths, sizes, hashes, and aggregate are hard-pinned outside the generated manifest; the legal files are scanned normally and the package guard compares size and SHA-256 for every canonical frontend corpus file, legal companion, runtime closure, manifest, and family adapter between a verified pre-pack snapshot and its produced temporary tarball. This rejects lifecycle mutation and coordinated legal-byte/manifest repinning rather than trusting path presence. Each verifier returns an immutable expected file map bound to manifest verification or independent commitments; capture checks each exact buffer against that map, so mutation in the verifier-to-capture interval fails. Expected entries also bind regular non-executable semantics across source reads and tar headers. Any execute bit fails, while secure non-executable modes such as `0600` remain valid without requiring exact `0644`. Canonical frontend preflight uses an 8 MiB per-file and 32 MiB aggregate budget, while runtime closure preflight uses the same 8 MiB per-file and 32 MiB aggregate budget before any whole-file read. ## Runtime Concepts | Reference surface | LitClaude target | Status | | --- | --- | --- | | prompt activation | `hooks/hooks.json`, `bin/litclaude-hook.js`, `commands/lit-loop.md`, `commands/lit-plan.md`, `commands/deep-interview.md`, `commands/start-work.md` | UserPromptSubmit detects workflow triggers and injects compact Claude Code guidance. | | comment-checker | `skills/comment-checker`, `PostToolUse` hook guidance | Claude-native comment hygiene skill plus post-edit reminder. | | lsp | `.lsp.json`, `.mcp.json`, `bin/litclaude-lsp-doctor.js`, `skills/lsp` | Lightweight Claude Code LSP/MCP helper rather than vendoring a full external package. | | rules | `skills/rules`, `SessionStart` (`source: compact`) hook behavior | Claude session rule guidance and compact cache-reset messaging. | | litgoal | `commands/litgoal.md`, `skills/litgoal`, `plugins/litclaude/lib/litgoal/`, `lit-*`, `start-work` goal guidance | Claude-native goal wording with evidence/ledger/quality-gate discipline and durable CLI state. | | litwork | `commands/lit-loop.md`, `commands/lit-plan.md`, `skills/lit-loop`, `skills/lit-plan` | Claude skills, commands, agents, workflow/worktree guidance, and hook injection. | | clarification interview | `commands/deep-interview.md`, `skills/deep-interview` | Socratic requirements mode with ambiguity scoring, persisted state, and a handoff to planning/execution. | | agents | `agents/*.md` | Planner, executor, verifier, reviewer, librarian, and QA runner roles. | | mcp | `.mcp.json`, `bin/litclaude-mcp.js` | Plugin-local stdio helper surface. | | install metadata | `bin/litclaude-ai.js`, `.claude-plugin/plugin.json`, local marketplace metadata | Installer writes Claude user plugin registry and LitClaude-owned local marketplace records. | | telemetry | No runtime telemetry | Intentionally omitted as network telemetry. LitClaude remains quiet/local by default. | ## Skill Corpus LitClaude includes the compatibility skill corpus: - `lit-burnoff-file` - `comment-checker` - `debugging` - `frontend-ui-ux` - `lsp` - `lit-code` - `refactor` - `lit-burnoff` - `review-work` - `rules` - `litgoal` LitClaude now ships those names or Claude-native equivalents, plus LitClaude workflow additions: - `start-work` - `review-work` - `litgoal` - `lit-loop` - `lit-plan` - `lit-handoff` - `lit-scientific-visualization` - `autoresearch` - `autoconference` - `wikify` Auxiliary Skill-discovery entries are part of the public inventory without adding new command routes: - `frontend-ui-ux` - `lit-commit` - `lsp-setup` - `visual-qa` ## Operator Contract The workflow documentation is intentionally local-first. It does not claim npm publication, remote marketplace registration, or public promotion. The publish boundary remains explicit user approval, and uninstall remains limited to LitClaude-managed state. Review work uses a 5-lane review: scope/diff verification, tests/evidence execution, package/payload and code quality, security/provenance, and real-surface/docs readiness. Manual-QA channels must leave artifacts and a cleanup receipt. Litgoal state is managed with: ```bash litclaude litgoal create-goals --brief "" --json litclaude litgoal record-evidence --criterion --status pass --json '{"artifact":"...","cleanup":"..."}' litclaude litgoal checkpoint --status active --note "" --json litclaude litgoal steer --kind scope --note "" --json ``` Current docs/test reads override stale state during audits. Prompt text is not executed or echoed, and malformed input should produce a controlled error. The original `lit-burnoff-file` command-facing name is preserved as an alias-style skill so older operator muscle memory still has a matching Claude surface. `lit-burnoff` remains the broader LitClaude cleanup skill. ## Auxiliary Packs The deeper workflow skills rely on files beyond `SKILL.md`. LitClaude now packages the high-value portable auxiliary packs: - `lit-code/references` - `lit-code/scripts` - `debugging/references` These packs carry language/runtime-specific guidance for TypeScript, Python, Rust, Go, debugging methodology, runtime probes, and tool usage. Non-Claude coordination examples in the debugging references were rewritten as Claude workflow/subagent pseudocode so they do not instruct Claude Code to call missing tools literally. ## Deliberate Omissions The following reference workflow content was audited but not copied into the runtime package: - Web/landing app: not part of the Claude Code plugin runtime. - GitHub Actions and npm publish workflows: this repository uses manual user publication. - Telemetry network component: omitted to keep LitClaude local and quiet. - Full component TypeScript packages for rules/comment-checker/LSP/litgoal: represented by Claude-native hook/skill/MCP/LSP surfaces where portable, not vendored wholesale. If any of those become desired runtime features, they should be introduced with their own tests and explicit user approval.