Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/lacolaco/claude-plugins/kb-ingestnpx skills add lacolaco/claude-plugins --skill kb-ingestgit clone --depth 1 https://github.com/lacolaco/claude-pluginsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/lacolaco/claude-plugins/kb-ingest)<a href="https://agentmods.dev/skills/lacolaco/claude-plugins/kb-ingest"><img src="https://agentmods.dev/badge/skills/lacolaco/claude-plugins/kb-ingest.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00375 | $0.01982 |
| Opus 5 | $0.00187 | $0.00991 |
| Sonnet 5 | $0.00075 | $0.00396 |
| Haiku 4.5 | $0.00038 | $0.00198 |
Grade A, and why
kb-ingest scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kb-ingest (ingest / deingest)
The KB lives at ~/.knowledge/. The storage schema (raw/wiki layout, OKF v0.1 conformance, page conventions, status notation, symlink spec) is in ~/.knowledge/CLAUDE.md — read it before starting.
The wiki/ layer is an OKF v0.1 bundle — every page you write or update must satisfy §9.1 (parseable YAML frontmatter + non-empty type).
Ingest
-
Read the schema (
~/.knowledge/CLAUDE.md). -
Collect intermediate-artifact files into
~/.knowledge/raw/: scan~/works/<project>/docs/{designs,reviews,plans}/and any other location holding AI-generated middle artifacts (acceptance-test records, retrospectives, conversation logs). Copy each into~/.knowledge/raw/<project>/<category>/<name>.mdif not already mirrored. Tracking state is irrelevant — commit-tracked files are mirrored too. Never delete the source as a shortcut and never skip the copy because "it's already committed in the repo". -
Read the target at manifest depth only (no full-source reads). Create/update
~/.knowledge/wiki/projects/<name>.md→ updatewiki/entities/andwiki/topics/for cross-cutting elements and add cross-links → register inwiki/index.md→ append towiki/log.mdunder the matching## YYYY-MM-DDH2 heading (or add a new heading newest-on-top). Cite the source raw path for every claim (raw/<project>/...for mirrored artifacts; anchor path —works/<path>or~/.claude/<path>— for source code, manifests, README, CLAUDE.md); mark unverified items as(guess). -
Write OKF frontmatter at the top of every wiki page (see schema for controlled
typevocabulary):--- type: Project Wiki # or Entity Wiki / Topic Wiki title: <human-readable name> description: <one-sentence summary> # recommended tags: [<keyword>, ...] # recommended timestamp: <ISO 8601 UTC> # this ingest's clock source_paths: # optional - works/<project>/<file or dir> source_commit: <git SHA> # optional but recommended --- -
Do not duplicate ground truth. Versions, package manager, presence of CLAUDE.md, etc. come from
package.json/ lockfile / filesystem at query time. The wiki holds only knowledge that cannot be re-derived from the primary source (purpose, cross-cutting relations, pitfalls, contradictions). Verify package manager from the actual lockfile — do not trust external tables. -
Large monorepos / framework forks such as
angularget a pointer page only. -
Mandatory post-write sanitize (textlint). After every wiki write/update (project page, entity/topic page,
index.md,log.md), run thememory-sanitizeskill'scheck.shon<changed wiki .md files>and reach 0 violations before moving to step 8. The check surfaces candidates — judge each: if a real violation, restructure the prose; if a false positive (heading-then-list:etc.), leave as-is and note why. Do not pass~/.knowledge/raw/*or the allowlist*.jsonfiles to check.sh — they are not prose. Restrict the argument list to the.mdfiles this ingest touched. -
Mandatory post-write tech-writing review. After step 7 reaches 0 violations, perform the
tech-writingskill review on every wiki.mdpage this ingest touched. Read the skill's normative sections in full and inspect each touched page for violations the mechanical lint cannot detect — paragraph structure, argumentative rigor, redundancy, performative phrasing, LLM-style filler. Fix violations or, for justified exceptions, leave a one-line reason. If the review modifies prose, re-run step 7'scheck.shto confirm textlint still passes. -
Mandatory post-write OKF check. Run
thekb-lintskill'skb-lint.sh`` and confirm the §0 conformance row reportsfrontmatter 欠落 0 / type 欠落 0 / 予約構造違反 0. Any non-zero count blocks declaring the ingest complete — fix the offending page (addtype:or restore the reserved file's required structure) and re-run. Declaring the ingest complete requires steps 7, 8, and 9 to have passed.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 73 lines · 375 tokens per session scan A 1c27f8196e4e
kb-ingest is a skill published in the GitHub repository lacolaco/claude-plugins (5 stars, last pushed 10d ago), licensed MIT. It adds 375 tokens to every session and 1,982 once invoked, about $0.0019 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
media-ingest
Ingest video, audio, PDF, book, screenshot, and GitHub repo content into the brain. Multi-format handling with entity extraction and backlink propagation. Covers video-ingest, youtube-ingest, and book-ingest subtypes.
mem0-oss-to-platform
Plan and then execute a migration of a project from the mem0 open-source / self-hosted SDK (the local Memory class) to the mem0 Platform / hosted / managed SDK (the MemoryClient class). Use this whenever a developer wants to move, switch, or migrate their mem0 usage off OSS/self-hosted to the hosted API — e.g.…
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…
agent-memory
../../../engineering/agent-memory/skills/agent-memory/SKILL.md.
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.
ha-data-stores
Map of Hope Agent's local data stores and safe read-only query workflow. Use when the user asks where Hope Agent stores data, wants to inspect sessions/messages/memory/logs/background jobs/knowledge indexes/settings, asks the model to query local app data, or debugging requires checking persisted state. Trigger…