codyswanngt/lisa-wiki-ingest
plugins/lisa-wiki/skills/lisa-wiki-ingest/SKILL.md
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.
$ install --global
skillsauthnpx skillsauth add codyswanngt/lisa lisa-wiki-ingestInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 27, 2026, 4:05 AM170.2s2 files scanned
SKILL.md
- 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-writeconnectors (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.
- Resolve the default remote branch —
gh repo view --json defaultBranchRef -q .defaultBranchRef.name, orgit remote show origin | sed -n 's/.*HEAD branch: //p', or theorigin/HEADsymbolic ref. If the repo has no remote, note "no remote — skipping branch sync" and proceed. - Fetch —
git fetch <remote>to update remote-tracking refs. - 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.
- On the default branch → fast-forward to
- 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
- 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. - 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. - Index: update
wiki/index.md. - Log: append a
wiki/log.mdentry (fixed operation vocabulary). - Verify:
git diff --check, secret/tenant/contamination scans, touched-file guard. - State: advance the connector's
wiki/state/<system>/*.jsoncursor — only now, after 1–5 pass. - Commit/PR: commit only the ingestion changes per
config.gitpolicy. 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-writeruns 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-rundoes 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
sourceRetentionandsensitivity; 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.
Related Skills
codyswanngt/lisa-wiki-onboard-me
documentation
VerifiedTrustedCommunity
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.
1SKILL.mdUpdated Jun 5, 2026
codyswanngt/lisa-wiki-migrate
documentation
VerifiedTrustedCommunity
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.
1SKILL.mdUpdated Jun 5, 2026
codyswanngt/lisa-wiki-lint
development
VerifiedTrustedCommunity
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.
1SKILL.mdUpdated Jun 5, 2026
codyswanngt/lisa-wiki-ingest
testing
VerifiedTrustedCommunity
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.
1SKILL.mdUpdated Jun 5, 2026