@codyswann/lisa 2.121.0 → 2.122.0

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-doctor/SKILL.md

@@ -0,0 +1,47 @@
+---
+name: lisa-wiki-doctor
+description: Verify that a wiki is correctly set up and (after migration) fully functional. Runs the deterministic checks plus functional smoke tests, writes a doctor report, and returns an overall verdict. Use as the final gate of /migrate, after /setup, or anytime to confirm wiki health. A repo is not "migrated" until its doctor verdict is READY.
+---
+
+# lisa-wiki-doctor
+
+The post-migration / post-setup verification checklist. Orchestrates the deterministic scripts and
+functional smoke tests, then writes `wiki/state/migration/doctor-report.json` and prints the blocking
+items.
+
+## Verdict
+Each item is `PASS | WARN | FAIL | SKIP`; overall is `READY | READY_WITH_WARNINGS | NOT_READY`. In
+hard-enforcement mode (`--migration` after phase 5) `WARN` on structure/integrity items is not
+allowed; earlier phases allow documented legacy warnings. Build checks (group G) are Lisa/release-only
+and `SKIP` in downstream repos.
+
+## Checklist
+- **A. Structure & config** (deterministic): `validate-config.mjs` + structure-manifest conformance;
+  required files present; `schemaVersion` set and rendered `kernelVersion` matches the installed
+  plugin; README mode recorded (never implicit `stub`).
+- **B. Integrity & safety** (deterministic): `lint-wiki.mjs` + `diff-guard.mjs`; index/log coverage;
+  state-order invariant; no broken/orphan links; secret/tenant/contamination/sensitivity scans;
+  memory sources project-scoped only (no global Codex memory / Chronicle).
+- **C. No-loss / parity**: pre/post manifest maps every legacy page/doc/command/ingest-path/role;
+  nothing dropped; docs absorption idempotent with zero dangling links; old README ingested before
+  rewrite; no leftover project-local machinery except generated overlays + role agents.
+- **D. Runtime surfaces** (both runtimes): canonical commands resolve on Claude; `$lisa-wiki-*` on
+  Codex; each `config.staff[]` role has its doc page + both subagents; MCP doctor passes or connector
+  cleanly disabled; external-write connectors skipped unless explicit intent.
+- **E. Functional smoke** (skill-orchestrated): targeted ingest of a small fixture exercises the full
+  ordered pipeline (in the migration branch or dry/scratch mode — **no extra PR**); `/query` returns
+  a cited answer; `/lint` clean (or expected phase warnings); `/onboard-me` completes read-mostly with
+  no PII; bare `/ingest --dry-run` lists non-external-write sources; `/doctor` rerun is idempotent.
+- **F. Mode-specific**: embedded self-ingest at HEAD; wrapper child-docs ingested-not-moved and not
+  staged; standalone/subdir path resolution; no-PR mode does not require PR creation.
+- **G. Git / CI / distribution** (deterministic): clean working tree (only intended changes); no
+  secrets/OAuth committed; PR/auto-merge policy honored; CI validator green if `--with-ci`; (Lisa
+  only) `bun run build:plugins` builds both artifacts and `check:plugins` passes.
+
+## Notes
+- Deterministic items are run by `scripts/verify-migration.mjs` (which emits the JSON report); this
+  skill adds the functional smoke tests and interprets phase-allowed warnings.
+- Read-mostly: the smoke tests must not open an extra PR or perform external writes.
+
+## Related
+`lisa-wiki-migrate`, `lisa-wiki-lint`, `lisa-wiki-setup`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-ingest/SKILL.md

@@ -0,0 +1,72 @@
+---
+name: lisa-wiki-ingest
+description: Ingest source material into the LLM Wiki. With an argument (URL, file path, or prompt) it ingests that one source; with no argument it runs a full ingest across every enabled non-external-write source. Routes to the right connector, then runs the ordered pipeline (source note → synthesis → index → log → verify → state → commit/PR). Use whenever new knowledge should enter the wiki.
+---
+
+# lisa-wiki-ingest
+
+The single entry point for getting knowledge into the wiki. It is a **router**: it never reimplements
+synthesis per source — it selects a connector to produce a sanitized source note, then the kernel
+performs the shared, ordered pipeline.
+
+## Modes
+- **Targeted:** `/ingest <url|file|prompt>` — ingest one source. Classify the input and pick the
+  matching connector.
+- **Full:** `/ingest` (no argument, or "do a full ingest") — iterate **every enabled connector whose
+  side-effect policy permits unattended ingest**: self + other registered projects' git/PR history,
+  project-scoped memory, roles, plus read-only registered sources (notion, jira, confluence,
+  quickbooks, …). `external-write` connectors (Slack OAuth, CRM writeback) are **skipped unless the
+  run includes explicit external-write intent**.
+- **Dry run:** `/ingest --dry-run` — list the sources a full ingest would run; perform no writes.
+
+## Before ingesting — sync the branch (once per run)
+
+Run this **once per ingest invocation, before any source is processed** (skip for `--dry-run`, which
+writes nothing). The point is to ingest on top of fresh state, never stale state.
+
+1. **Resolve the default remote branch** — `gh repo view --json defaultBranchRef -q .defaultBranchRef.name`,
+   or `git remote show origin | sed -n 's/.*HEAD branch: //p'`, or the `origin/HEAD` symbolic ref. If
+   the repo has no remote, note "no remote — skipping branch sync" and proceed.
+2. **Fetch** — `git fetch <remote>` to update remote-tracking refs.
+3. **Bring the working branch up to date with the default remote branch** so the ingest lands on
+   current state:
+   - On the default branch → fast-forward to `<remote>/<default>` (`git pull --ff-only`).
+   - On a non-default branch → merge or rebase `<remote>/<default>` in (per the project's convention)
+     so the branch is not behind the default.
+4. If the sync cannot complete cleanly (merge conflict, diverged history, or a dirty tree that would
+   conflict), **stop and surface it** rather than ingesting on top of stale or conflicted state — the
+   human resolves and re-runs. **Never discard unrelated working-tree changes** to force a sync.
+
+## The ordered pipeline (per source) — never reorder
+1. **Connector** validates (tenant guard + auth), reads its state cursor (first-run vs incremental),
+   fetches read-only, and writes a sanitized **source note** under `wiki/sources/<system>/` plus run
+   metadata. A connector writes *only* its source note + metadata — never synthesis/index/log/state.
+2. **Synthesis** (kernel): distill durable knowledge into the appropriate category pages, with
+   citations back to the source note. Weak/uncertain material → `wiki/open-questions/`, never asserted.
+3. **Index**: update `wiki/index.md`.
+4. **Log**: append a `wiki/log.md` entry (fixed operation vocabulary).
+5. **Verify**: `git diff --check`, secret/tenant/contamination scans, touched-file guard.
+6. **State**: advance the connector's `wiki/state/<system>/*.json` cursor — only now, after 1–5 pass.
+7. **Commit/PR**: commit only the ingestion changes per `config.git` policy. If the ingest started on
+   the default branch, create a dedicated ingestion branch first — never commit ingestion straight to
+   the default. Push the branch and **open a PR targeting the default remote branch** (via the host's
+   PR mechanism — e.g. `gh pr create --base <default>`), then **enable auto-merge when possible**
+   (`gh pr merge --auto`, or the host's equivalent). `external-write` runs and sensitive content are
+   the exception — open the PR **without** auto-merge so a human reviews them before it lands. If
+   auto-merge cannot be enabled (the host doesn't support it, or branch protection forbids it), leave
+   the PR open and note that a human must merge.
+
+## Rules
+- **Bookend every ingest with git hygiene:** sync the branch with the default remote branch *before*
+  writing (see "Before ingesting"), and *after* a successful ingest open a PR to the default remote
+  branch with auto-merge enabled when possible — never auto-merging `external-write`/sensitive runs.
+  `--dry-run` does neither (it writes nothing).
+- Source-note-before-synthesis; state advanced **only** after verification.
+- Project-scoped only; memory ingestion never touches global/Codex-global stores.
+- Respect `sourceRetention` and `sensitivity`; do not invent facts.
+- Connector execution and the connector contract are detailed in the connector skills (M2+); this
+  router defines and enforces the ordering and side-effect rules above.
+
+## Related
+`lisa-wiki-add-ingest` (scaffold a custom front-door that chains here), `lisa-wiki-query`,
+`lisa-wiki-lint`, `lisa-wiki-doctor`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-lint/SKILL.md

@@ -0,0 +1,32 @@
+---
+name: lisa-wiki-lint
+description: Health-check the LLM Wiki. Reports orphan pages, contradictions, stale claims, broken internal links, missing index/log coverage, structure-manifest violations, and secret/tenant leaks. Use periodically or before hardening a wiki. Read-only — it reports findings, it does not fix them.
+---
+
+# lisa-wiki-lint
+
+Run the wiki's integrity checks and report findings by severity. Lint is **read-only**: it diagnoses,
+it does not repair (use `/ingest`, `/setup`, or `/migrate` to fix).
+
+## What it checks (deterministic core via `scripts/lint-wiki.mjs`)
+- **Frontmatter** present/valid where required (per `schema/page-frontmatter.schema.json`).
+- **Index coverage**: every page appears in `wiki/index.md`; no index rows point at missing pages.
+- **Log coverage**: material changes have `wiki/log.md` entries.
+- **Links**: no broken internal links; no orphan pages (unreferenced and unlinked).
+- **State ordering**: no state cursor advanced without its source notes + synthesis + index + log.
+- **Structure**: files conform to `schema/wiki-structure.schema.json` (canonical locations).
+- **Safety**: secret patterns, tenant/contamination terms, stray binaries, child-repo contents
+  staged in wrapper mode.
+
+## What it checks (LLM-assisted)
+- **Contradictions** between claims; **stale** claims (older than the configured staleness window);
+  **coverage gaps** vs the wiki's stated `purpose`.
+
+## Output
+A report grouped by severity, with each item marked `PASS | WARN | FAIL`. During phased migration,
+legacy issues are reported as `WARN`; in hard-enforcement mode structural/integrity issues are `FAIL`.
+Read-only: lint prints the report and does not modify the wiki (it does not write a `LINT` log entry).
+
+## Related
+`lisa-wiki-doctor` (broader post-migration verification that runs lint among other checks),
+`lisa-wiki-ingest`, `lisa-wiki-setup`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-migrate/SKILL.md

@@ -0,0 +1,43 @@
+---
+name: lisa-wiki-migrate
+description: Migrate an existing, hand-rolled wiki implementation onto the lisa-wiki kernel — phased and compatibility-first, with a strict no-loss guarantee. Use when adopting lisa-wiki in a repo that already has its own wiki/, ingest skills, docs, or roles. Renaming things into the canonical shape is fine; losing functionality or data is not. Ends by running /doctor.
+---
+
+# lisa-wiki-migrate
+
+Move a repo's bespoke wiki onto the kernel without breaking it or producing one giant diff. Each
+phase is a reviewable change; nothing legacy is deleted until its replacement is parity-checked.
+
+## Migration invariant — rename, never lose
+Any unique command, ingest path, or role is migrated **as though created via** `/add-ingest`
+(front-door skill) or `/add-role` (subagent); wiki-ingest aliases (`*-wiki-ingest`, custom `/ingest`
+routers, …) are renamed to the canonical `/ingest`. **Loss of functionality or data is not
+acceptable** — for each migrated artifact, run the old path and the new path and diff the output
+before deleting the old one.
+
+## Phases (per repo; one reviewable change each)
+- **0 — Inventory.** Profile the repo: mode, root, categories, frontmatter coverage, log/index
+  format, source dirs, state files, connectors, runtime surfaces, MCP config, scripts, PR policy.
+  Write a pre/post manifest mapping every legacy page, source note, doc, command, ingest path, role.
+- **1 — Adopt kernel (no content rewrite).** Add `wiki/lisa-wiki.config.json`, render the contract
+  snapshot, add `wiki/state/README.md`, run validators in **warning** mode.
+- **1b — Documentation absorption & structure.** `scripts/absorb-docs.mjs` moves the host repo's own
+  docs (`docs/`, `specs/`, top-level docs) into `wiki/documentation/`, ingests them, and conforms to
+  the structure manifest; `scripts/rewrite-refs.mjs` rewrites every internal link/citation/index
+  entry (zero dangling links). Keep-in-place files stay at conventional paths. **Ask** the README
+  mode (default `rich`; `stub` only on explicit choice). Wrapper mode moves only host docs, never
+  child-project docs. Idempotent (source fingerprints); parity-checked.
+- **2 — Runtime consolidation.** Replace duplicated local wiki skills with the plugin's; bespoke ones
+  become `/add-ingest` front-doors or `/add-role` subagents. Claude commands become facades.
+- **3 — Connector consolidation.** Centralize shared connectors (e.g. Slack); keep org-specific ones
+  as contrib/overlay.
+- **4 — Normalize by touch.** New/touched pages get frontmatter; new logs use the canonical table;
+  legacy stays parseable until deliberately migrated.
+- **5 — Hard enforcement.** Flip validators to hard-fail (and enable CI) once the baseline is clean.
+
+## Finish
+Run `lisa-wiki-doctor --migration`. The repo is not considered migrated until the verdict is `READY`
+(or a human-approved `READY_WITH_WARNINGS`). Roll back by reverting the migration change.
+
+## Related
+`lisa-wiki-doctor`, `lisa-wiki-setup`, `lisa-wiki-add-ingest`, `lisa-wiki-add-role`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-onboard-me/SKILL.md

@@ -0,0 +1,33 @@
+---
+name: lisa-wiki-onboard-me
+description: Onboard a user to the project via its LLM Wiki. Interviews the user about themselves in relation to the project, captures that to project-scoped memory only, then gives a guided tour of what the project is and sample questions they can ask. Use when someone is new to the project or asks to be onboarded. Read-mostly — it does not open PRs or write PII into the wiki.
+---
+
+# lisa-wiki-onboard-me
+
+Bidirectional onboarding: learn who the user is *in relation to this project*, then teach them what
+the project is and how to use the wiki. The minimal project README points here.
+
+## Workflow
+1. **Interview** (brief): the user's role, goals, and what they intend to do in this project.
+2. **Capture** to **project-scoped memory only** — Claude's per-project memory; on Codex only if a
+   project-scoped memory store exists. If none exists, keep it session-local. **Never** write the
+   capture into global memory and **never** commit PII into the wiki.
+3. **Tour:** summarize the project from `wiki/start-here.md` + the wiki's `purpose`, show the folder
+   map (categories), and point at the key entry pages.
+4. **Sample questions:** offer a handful of `/query` prompts tuned to the user's role to get them
+   productive immediately.
+
+## Modes
+- Default: **read-mostly, no PR.**
+- `--save-memory`: persist the capture to project-scoped memory.
+- `--write-audience-note`: write a sanitized, **non-PII** audience/role note into the wiki via the
+  normal PR flow — only when `onboarding.allowAudienceNote: true` in config.
+
+## Rules
+- Project-scoped only; global Codex memory and the Chronicle store are never used.
+- Do not auto-create a staff role; offer `/add-role` only if the user is taking on a durable project
+  function (not merely onboarding).
+
+## Related
+`lisa-wiki-query`, `lisa-wiki-usage`, `lisa-wiki-add-role`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-query/SKILL.md

@@ -0,0 +1,30 @@
+---
+name: lisa-wiki-query
+description: Answer a question from the LLM Wiki with citations. Reads the index, drills into relevant pages, and synthesizes a cited answer. Read-only by default; only files new synthesis back into the wiki when the user explicitly asks. Use when someone asks a question the wiki should be able to answer, or wants to explore what the wiki knows.
+---
+
+# lisa-wiki-query
+
+Answer from the wiki, with citations, without changing it (by default).
+
+## Workflow
+1. Read `wiki/index.md` to locate candidate pages; consult `wiki/start-here.md` for orientation.
+2. Drill into the relevant synthesis pages and their cited source notes.
+3. Synthesize an answer. **Every claim cites its wiki page and/or source note.** If the wiki does not
+   support an answer, say so plainly rather than inventing one; suggest an `/ingest` that would fill
+   the gap. Only record the gap in `wiki/open-questions/` if the user explicitly asks (that is a
+   writeback — see step 4); never write it silently.
+4. **Writeback (opt-in only):** if — and only if — the user explicitly asks to persist new synthesis
+   discovered during the query, route it through the ingest pipeline so provenance/index/log/state
+   stay consistent, and log it as a `QUERY` operation. Never write back silently.
+
+## Rules
+- **Read-only by default.** No page edits, index/log changes, state advancement, or PRs unless the
+  user explicitly requests writeback.
+- Prefer the wiki's own content over outside knowledge; distinguish wiki-sourced facts from your own
+  reasoning.
+- Surface contradictions or stale claims you encounter as `/lint`-style findings rather than
+  silently picking one.
+
+## Related
+`lisa-wiki-ingest`, `lisa-wiki-lint`, `lisa-wiki-usage`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-setup/SKILL.md

@@ -0,0 +1,75 @@
+---
+name: lisa-wiki-setup
+description: Scaffold, repair, verify, or upgrade a project's LLM Wiki from its config. Use when setting up the wiki in a new repo, fixing a broken/incomplete structure, or upgrading to a newer kernel version. Asks the wiki's purpose and README mode, renders the contract snapshot, scaffolds the canonical folders, and seeds the staff roster. Idempotent and non-destructive.
+---
+
+# lisa-wiki-setup
+
+Bring a repo's `wiki/` into conformance with the canonical structure from
+`wiki/lisa-wiki.config.json`. Safe to re-run: it creates what is missing and repairs drift, but
+never overwrites human-authored content.
+
+## When to use
+- First-time setup of the wiki in a repo.
+- Repairing a wiki that fails `/doctor` structural checks.
+- Upgrading after a new `lisa-wiki` release (`--upgrade`): re-render the contract snapshot and run a
+  compatibility report before writing.
+- `--with-ci`: also install the optional GitHub Action validator (`ci/lisa-wiki-validate.yml`).
+
+## Workflow
+1. **Config.** Read `wiki/lisa-wiki.config.json`, or create it interactively. Ask for: `org`,
+   `displayName`, **`purpose`** (one paragraph — what this wiki is for), `mode`
+   (`embedded|wrapper|standalone|subdir`), `categories`, source layout, connectors, sensitivity,
+   `sourceRetention`, and the **README mode** (`rich` default | `stub` | `preserve` — always ask;
+   never select `stub` implicitly). If `staff` is absent, seed the **standard roster** (below) as the
+   default — every wiki gets the standard operating team unless the user opts out or edits it. Union
+   each role's owned categories into `config.categories`. Validate with `scripts/validate-config.mjs`.
+2. **Structure.** Scaffold the canonical tree per `schema/wiki-structure.schema.json`:
+   `wiki/{index.md, log.md, start-here.md, schema/, sources/, state/, staff/, <category dirs>}`.
+   Create only what is missing.
+3. **Contract.** Render `wiki/schema/llm-wiki-contract.md` from the plugin templates + config via
+   `scripts/render-contract.mjs`, stamping the `kernelVersion`. This snapshot keeps the wiki
+   self-describing without the plugin installed.
+4. **Gitignore.** Merge the lisa-wiki gitignore block into the project's `.gitignore` via
+   `scripts/ensure-gitignore.mjs`. The block (delimited by `# BEGIN: AI GUARDRAILS WIKI` /
+   `# END: AI GUARDRAILS WIKI`) covers transient per-session worktrees and Lisa backup snapshots
+   (`.claude/worktrees/`, `.codex/worktrees/`, `.lisabak/`). Idempotent: re-running produces no
+   diff once the block is present. The block coexists with the base lisa plugin's
+   `# BEGIN: AI GUARDRAILS` block — both can be installed without overwriting each other because
+   the copy-contents strategy keys on the marker suffix. Wiki-wrapper repos (mode `wrapper` /
+   `standalone`) typically don't enable the base lisa plugin, so this step is the only path by
+   which they get the worktree-ignore patterns.
+5. **Pointers.** Ensure `AGENTS.md` / `CLAUDE.md` point at the contract + plugin (thin pointers only).
+6. **Staff.** For each `config.staff[]` entry (the standard roster by default), generate the role's
+   `wiki/staff/<role>.md` page and its dual-runtime subagents by delegating to `lisa-wiki-add-role`
+   (running the subagents is out of scope).
+7. **README.** Apply the chosen README mode (ingest the old README first; `rich` keeps install/usage +
+   adds the onboarding line; `stub` is the minimal pointer; `preserve` leaves it).
+8. **Verify.** Run `lisa-wiki-doctor` and report the verdict + any blocking items.
+
+## Standard roster
+The default operating team seeded into `config.staff[]` for every new wiki (Chief of Staff plus six
+domain agents). Each role becomes a `wiki/staff/<id>.md` page and a dual-runtime subagent; the human
+owner talks to Chief, who routes to the others. Owned categories are unioned into `config.categories`.
+
+| id | role | owns (categories) | sensitivity |
+|---|---|---|---|
+| `chief` | Chief of Staff | `projects`, `decisions`, `playbooks`, `open-questions` | confidential |
+| `sally` | Sales | `sales` | internal |
+| `mark` | Marketing | `marketing` | internal |
+| `felix` | Finance | `finance` | confidential |
+| `casey` | Customer Success | `customers` | internal |
+| `parker` | People | `people` | confidential |
+| `lex` | Legal & Compliance | `legal` | confidential |
+
+Projects may add, remove, or rename roles afterward; the standard roster is the starting point, not a
+constraint.
+
+## Rules
+- Idempotent; re-running produces no spurious changes.
+- Never overwrite human content; only create/repair structure and the rendered snapshot. The README is
+  rewritten only after its old content has been ingested, and only per the chosen `readme.mode`.
+- Project-scoped only; never stage secrets/OAuth artifacts; honor `mode` safety (wrapper/standalone).
+
+## Related
+`lisa-wiki-add-role`, `lisa-wiki-doctor`, `lisa-wiki-migrate`, `lisa-wiki-usage`.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-setup-automations/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: lisa-wiki-setup-automations
+description: "Set up the recurring LLM Wiki ingest automation on the local workstation using the CURRENT runtime's native scheduler — Codex automations (the native automations / automation_update mechanism) or, on Claude, /schedule. This skill is a declarative specification: it states WHICH automation to create, how often, and with which command; it does not template schedule files or run scheduling code itself — the runtime's native automation mechanism does the creating. Creates one automation: wiki-ingest, a full /lisa-wiki:ingest cycle, once a day by default (override with cadence). Named and scoped lisa-wiki-auto-<project>-* so it never collides with or clobbers the base /setup-automations set. The wiki counterpart of /lisa:setup-automations — it lives here because the wiki plugin is standalone and installable without the base plugin. Tear down with /lisa-wiki:tear-down-automations."
+allowed-tools: ["Skill", "Bash", "Read"]
+---
+
+# Set up LLM Wiki automations: $ARGUMENTS
+
+This skill is a **specification, not a script.** It tells the current runtime which recurring wiki
+automation to create, on what cadence, and with which command — and the runtime creates it with its
+**native** scheduling mechanism. Do **not** hand-template schedule files or write shell to create
+them; invoke the runtime's automation tool with the spec below.
+
+It is the wiki counterpart of the base `/lisa:setup-automations`. It is a **separate** skill because
+the wiki plugin (`lisa-wiki`) is standalone — it can be installed without the base Lisa plugin, in
+which case `/lisa:setup-automations` is not present to schedule ingest. The two skills are
+independent and use disjoint name prefixes, so running both is safe.
+
+## Runtime scheduler (branch on the current runtime)
+
+- **Codex** → create a Codex **automation** via the native automations mechanism (prefer the
+  `automation_update` tool over hand-writing `~/.codex/automations/<id>/automation.toml`; the TOML is
+  only its backing store). Set the execution environment to **local** so it runs on this
+  workstation. Scope it to a durable project automation checkout, not a transient task worktree: use
+  `${CODEX_HOME:-~/.codex}/worktrees/<project>-automation-main` when available, create or refresh
+  that checkout from the project's `origin` remote if needed, and verify `git -C <cwd> rev-parse
+  --is-inside-work-tree --is-bare-repository` reports `true` then `false` before saving the
+  automation. Do not point recurring automations at hashed scratch worktrees or a checkout whose Git
+  metadata is broken.
+- **Claude** → use **`/schedule`** to create a local recurring routine.
+- **Other runtimes** → use the runtime's native recurring-task mechanism. If the runtime has none,
+  state that scheduling is unavailable and stop.
+
+## Parameters
+
+- `cadence` (default **daily**) — how often the full ingest runs. Accepts `daily`, `weekly`, or an
+  `every-<n>-hours` form. Map to a Codex `rrule`: daily → `FREQ=DAILY;INTERVAL=1`; weekly →
+  `FREQ=WEEKLY;INTERVAL=1`; every N hours → `FREQ=HOURLY;INTERVAL=<n>`. On Claude, pass the
+  equivalent `/schedule` cadence. Default is **daily** (`FREQ=DAILY;INTERVAL=1`).
+
+## The automation to create
+
+The automation runs **one cycle** of the full wiki ingest and respects that command's own confirmation
+and commit/PR policy (never ask before running; run a full ingest across every enabled
+non-external-write source; commit/PR per the ingest skill's bookends; report the cycle summary).
+Before running the ingest, the automation must sync its checkout: fetch the default remote branch and
+rebase the current automation branch onto it (for the common GitHub case, `origin/main`). If the
+checkout is already on the default branch, fast-forward/rebase it to the remote default. If the
+rebase has conflicts or the working tree is dirty in a way the automation did not create, abort the
+rebase and report the blocker instead of ingesting from stale code.
+
+| Automation | Command it runs | Cadence |
+|---|---|---|
+| **wiki-ingest** | `/lisa-wiki:ingest` (no argument → full ingest across all enabled sources) | once a day (or `cadence`) |
+
+**Naming + scope (so teardown is precise).** Name the automation with the stable prefix
+`lisa-wiki-auto-<project>-` (i.e. `lisa-wiki-auto-<project>-ingest`), where `<project>` identifies
+this repo, and scope each Codex automation to the durable project automation checkout described
+above. This prefix is deliberately distinct from the base `lisa-auto-<project>-` set so
+`/lisa-wiki:tear-down-automations` removes exactly this automation and never touches the base
+automations or any other project's. Use a project identifier stable across runs and distinct from
+other repos (qualify it, e.g. with the owner — don't rely on a bare repo basename that could
+collide).
+
+**Idempotent.** Re-running this skill updates the existing `lisa-wiki-auto-<project>-ingest`
+automation in place (same name) rather than creating a duplicate.
+
+## Conditions / guards
+
+- Create the automation only when this repo actually has an LLM Wiki — i.e. `wiki/` exists with a
+  `wiki/lisa-wiki.config.json`. If there is no configured wiki, **stop and report** that the wiki
+  must be set up first (run `/lisa-wiki:setup`); do not schedule ingest against a non-existent wiki.
+- If the runtime has no native scheduler, stop and report what's missing rather than guessing.
+- For Codex, if the durable checkout cannot be created, fetched, or verified as a non-bare Git work
+  tree, stop and report the checkout problem instead of creating an automation that will fail later.
+
+## Report
+
+List the automation created or updated (name, the command it runs, the resolved cadence), or report
+that it was skipped and why (no configured wiki / no runtime scheduler).

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-status/SKILL.md

@@ -0,0 +1,47 @@
+---
+name: lisa-wiki-status
+description: Report Lisa wiki source freshness across enabled ingestion connectors. Reads wiki config, log, source notes, and state cursors, then renders a concise read-only freshness table with skipped/blocker reasons and exact next actions. Use when maintainers ask whether the wiki sources are fresh, stale, skipped, blocked, or need targeted ingest.
+allowed-tools: ["Bash", "Read"]
+---
+
+# lisa-wiki-status
+
+Render the Lisa wiki source freshness report without changing the repository.
+
+## Scope
+
+This skill answers a narrower question than `lisa-wiki-doctor` or `lisa-wiki-lint`: whether each enabled, non-external-write ingestion connector has current source evidence and what the operator should do next.
+
+Read these repository files as the source of truth:
+
+- `wiki/lisa-wiki.config.json` for enabled connectors and wiki root.
+- `wiki/log.md` for latest ingest, skip, and blocker notes.
+- `wiki/sources/**` for reader-safe source notes.
+- `wiki/state/**` for connector state cursors and source-note references.
+
+## Workflow
+
+1. Resolve the wiki root from `wiki/lisa-wiki.config.json`, unless `--wiki` or `--config` is passed.
+2. Run the bundled deterministic renderer:
+
+   ```bash
+   node plugins/lisa-wiki/scripts/wiki-status.mjs $ARGUMENTS
+   ```
+
+   If running from the source plugin tree before distribution, use `plugins/src/wiki/scripts/wiki-status.mjs`.
+3. Report the rendered connector table. Preserve connector verdicts exactly:
+   `fresh`, `stale`, `never_ingested`, `skipped`, or `blocked`.
+4. Include the evidence paths, last observed date, skip/blocker reason when present, and exact next action for each non-fresh connector.
+5. Mention `lisa-wiki-lint` only as a separate integrity follow-up. Do not conflate freshness with broken links, stale claims, or structure linting.
+
+## Rules
+
+- **Read-only.** Do not ingest, edit wiki pages, edit config, advance state, branch, commit, push, open PRs, or run repair flows.
+- Do **not** ask for confirmation before rendering. This status command has no write side effects.
+- Do **not** use global Codex memory or non-project memory as fresh wiki evidence. Only project-scoped source notes, state, and log entries count.
+- If the wiki/config is missing, surface the renderer output or error directly and stop.
+- If a connector is skipped because project-scoped memory is unavailable, preserve that reason and treat accepting the expected skip as a valid next action.
+
+## Related
+
+`lisa-wiki-ingest` refreshes sources, `lisa-wiki-lint` checks wiki integrity, and `lisa-wiki-doctor` verifies setup/readiness.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-tear-down-automations/SKILL.md

@@ -0,0 +1,34 @@
+---
+name: lisa-wiki-tear-down-automations
+description: "Remove the recurring LLM Wiki ingest automation that /lisa-wiki:setup-automations created for this project (the lisa-wiki-auto-<project>-* set) using the CURRENT runtime's native scheduler — Codex automations or, on Claude, /schedule. This skill is a declarative specification: it identifies WHICH automation to remove; it does not run teardown scripts. Removes only this project's wiki automations — never the base /setup-automations set, other projects' automations, or non-Lisa ones. The inverse of /lisa-wiki:setup-automations."
+allowed-tools: ["Skill", "Bash", "Read"]
+---
+
+# Tear down LLM Wiki automations: $ARGUMENTS
+
+This skill is a **specification, not a script.** It tells the current runtime which recurring wiki
+automation to remove — the one `/lisa-wiki:setup-automations` created for THIS project — and the
+runtime removes it with its **native** scheduling mechanism.
+
+## Runtime scheduler (branch on the current runtime)
+
+- **Codex** → list Codex automations and delete the `lisa-wiki-auto-<project>-*` set via the native
+  automations mechanism (prefer the native delete over hand-removing
+  `~/.codex/automations/<id>/`, which is only the backing store).
+- **Claude** → use **`/schedule`** to list and remove the matching recurring routine.
+- **Other runtimes** → use the runtime's native recurring-task mechanism. If it has none, state that
+  and stop.
+
+## Scope (remove only what setup created)
+
+- Remove the wiki automation `/lisa-wiki:setup-automations` creates for the current project, matched
+  by the stable `lisa-wiki-auto-<project>-` name prefix: `wiki-ingest`.
+- **Never** remove the base `lisa-auto-<project>-*` automations, automations for a different project,
+  or any non-Lisa automation (e.g. unrelated crawlers/ingestors). Match strictly on the
+  `lisa-wiki-auto-<project>-` prefix for THIS project; when in doubt about an automation's ownership,
+  leave it and report it rather than deleting it.
+- **Idempotent** — an automation that is already absent is a no-op, not an error.
+
+## Report
+
+List the automation removed, or report it as already absent.

package/plugins/lisa-wiki-cursor/skills/lisa-wiki-usage/SKILL.md

@@ -0,0 +1,50 @@
+---
+name: lisa-wiki-usage
+description: Explain how to browse, query, and contribute to this project's LLM Wiki. Use when a user asks how the wiki works, where knowledge lives, how to find something, or how to add to it — the read/navigation path, not an ingestion or write workflow.
+---
+
+# lisa-wiki-usage
+
+Orient a human or agent to the project's LLM Wiki and point them at the right entry points and
+canonical commands. This skill is **read-only guidance** — it never writes wiki pages, advances
+state, or opens PRs.
+
+## What the wiki is
+
+A git-native markdown knowledge base under `wiki/`, maintained by the `lisa-wiki` plugin. It follows
+the three-layer model: immutable **raw sources** (`wiki/sources/`), the LLM-owned **synthesis** layer
+(category directories like `concepts/`, `entities/`, `architecture/`, …), and the **schema** that
+governs it (`wiki/schema/llm-wiki-contract.md`, rendered from the project's `wiki/lisa-wiki.config.json`).
+
+## Entry points (read these first)
+
+1. `wiki/start-here.md` — orientation and the wiki's stated purpose.
+2. `wiki/index.md` — the navigation map of every page (grouped by category).
+3. `wiki/schema/llm-wiki-contract.md` — the rules this wiki follows.
+4. `wiki/log.md` — the append-only history of every ingestion/maintenance operation.
+
+## How to do things
+
+On **Claude** these are slash commands; on **Codex** invoke the same skills from the app slash list
+(e.g. `$lisa-wiki-query`).
+
+- **Find or answer something:** `/query "<your question>"` (Codex: `$lisa-wiki-query`) — read-only by
+  default; returns a cited answer drawn from the wiki. Or browse `wiki/index.md` manually.
+- **Get oriented as a new user:** `/onboard-me` (Codex: `$lisa-wiki-onboard-me`) — a guided tour plus
+  sample questions tuned to your role.
+- **Add knowledge:** `/ingest <url|file|prompt>` (Codex: `$lisa-wiki-ingest`) for a single source, or
+  `/ingest` with no argument for a full ingest across all enabled **non-external-write** sources
+  (external-write sources like Slack OAuth require explicit intent). Do not hand-edit synthesis pages
+  to add facts — route them through `/ingest` so provenance, the index, and the log stay consistent.
+- **Check the wiki's health:** `/lint` (Codex: `$lisa-wiki-lint`) — orphans, contradictions, stale
+  claims, broken links.
+
+## Citations & trust
+
+Every synthesized claim should cite a source note (e.g. `Source: wiki/sources/<system>/<note>.md`).
+If you find an uncited or contradictory claim, that is a `/lint` finding — surface it rather than
+trusting it. Weak or unverified material belongs in `wiki/open-questions/`, not stated as fact.
+
+## Related skills
+
+`lisa-wiki-ingest`, `lisa-wiki-query`, `lisa-wiki-lint`, `lisa-wiki-onboard-me`, `lisa-wiki-setup`.

package/plugins/lisa-wiki-cursor/templates/agents/role-agent.claude.md

@@ -0,0 +1,16 @@
+---
+name: {{id}}
+description: {{roleDescription}}
+---
+
+You are **{{role}}** for {{org}} — the domain expert for {{expertise}}.
+
+Your knowledge lives in this project's LLM Wiki under: {{ownedPaths}}.
+
+Operating rules:
+- **Query the wiki first.** It is your source of truth — do not rely on stale or outside memory.
+  Use the `lisa-wiki-query` skill (`/query`) before answering.
+- **Contribute via ingestion.** Add new knowledge with `lisa-wiki-ingest` (`/ingest`) so provenance,
+  the index, the log, and state stay consistent. Never hand-edit synthesis pages to add facts.
+- **Stay in your lane.** Work within your owned domain; defer other domains to their roles.
+- **Respect sensitivity ({{sensitivity}})** and never expose secrets or out-of-scope material.

package/plugins/lisa-wiki-cursor/templates/agents/role-agent.codex.toml

@@ -0,0 +1,15 @@
+name = "{{id}}"
+description = "{{roleDescription}}"
+developer_instructions = """
+You are {{role}} for {{org}} — the domain expert for {{expertise}}.
+
+Your knowledge lives in this project's LLM Wiki under: {{ownedPaths}}.
+
+Operating rules:
+- Query the wiki first (use the lisa-wiki-query skill); it is your source of truth — do not rely on
+  stale or outside memory.
+- Contribute new knowledge via the lisa-wiki-ingest skill so provenance, index, log, and state stay
+  consistent; never hand-edit synthesis pages to add facts.
+- Stay within your owned domain; defer other domains to their roles.
+- Respect sensitivity ({{sensitivity}}); never expose secrets or out-of-scope material.
+"""

package/plugins/lisa-wiki-cursor/templates/index.md

@@ -0,0 +1,17 @@
+# {{displayName}} — Index
+
+> The navigation map of this wiki. Maintained on every ingestion. One table per category;
+> rows are `| Page | Summary | Updated |`. Read this first when looking for anything.
+
+## Orientation
+| Page | Summary | Updated |
+|---|---|---|
+| [start-here](start-here.md) | Orientation and the wiki's purpose | {{generatedDate}} |
+| [LLM Wiki Contract](schema/llm-wiki-contract.md) | The rules this wiki follows | {{generatedDate}} |
+
+<!-- Category sections (concepts, entities, …) are added/maintained by ingestion. -->
+
+## Sources
+| Source notes | Latest |
+|---|---|
+| `sources/` | (none yet) |