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/tessaryai/plugins/manage-knowledgenpx skills add tessaryai/plugins --skill manage-knowledgegit clone --depth 1 https://github.com/tessaryai/pluginsWhat 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 | $0.00065 | $0.00848 |
| Opus 5 | $0.00032 | $0.00424 |
| Sonnet 5 | $0.00013 | $0.00170 |
| Haiku 4.5 | $0.00006 | $0.00085 |
Grade A, and why
manage-knowledge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
manage-knowledge
Internal crew primitive — dispatched by
/crew:run. You are running because the orchestrator selected this as one step of a larger workflow; carry out the work below. This skill is not meant to be invoked on its own — user requests go to/crew:run.
You curate an evergreen knowledge base of durable decisions and gotchas — the "why"
behind the code that source and user docs don't capture. This complements update-docs
(which keeps user/developer docs accurate); here you distill decisions and lessons.
Argument is optional: a PR number captures from that PR; with no argument, do a sweep over recent merged PRs that aren't yet represented.
0. Load config and mode
python3 "${CLAUDE_PLUGIN_ROOT}/lib/load_config.py"
Use knowledge.dir (default docs/knowledge), labels.agent_pr, and ledger.dir.
Then read ${CLAUDE_PLUGIN_ROOT}/reference/work-model.md and resolve the mode before any
gh call.
1. Gather source material
- GitHub — PR mode:
gh pr view <N> --comments+gh pr diff <N>— read the description, the review thread, and the change itself. - GitHub — sweep mode:
gh pr list --state merged --limit 20 --json number,title,mergedAtand pick recent ones not yet referenced inknowledge.dir. - Local mode: read the task's
decision.mdandreview.mdfrom the ledger plus the branch's history (git log/git diff <base>...<branch>in the worktree).
2. Extract durable knowledge
Pull out only durable, reusable items — things a future engineer would want to know:
- Decisions — "we chose X over Y because Z" (with the trade-off and constraints).
- Patterns / conventions — a new approach the team adopted and why.
- Gotchas — non-obvious pitfalls, footguns, or constraints discovered the hard way.
Skip the ephemeral (one-off bug fixes with no lasting lesson, routine changes). If a PR yields nothing durable, record nothing for it.
3. Dedupe and write
- Read existing entries under
knowledge.dir. If an item is already covered, update that entry rather than adding a duplicate; only add genuinely new knowledge. - Append/update a dated entry in
knowledge.dir/decisions.md(create the file and a briefknowledge.dir/README.mdtopic index if absent), linking the source PR:
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.
- 2d ago First seen · 79 lines · 65 tokens per session scan A 263f358281b0
manage-knowledge is a skill published in the GitHub repository tessaryai/plugins (3 stars, last pushed 15d ago), licensed MIT. It adds 65 tokens to every session and 848 once invoked, about $0.0003 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
ghost-decode
Use when a video hides text in moving dots or noise — "ghost font" clips, motion-defined text, random-dot kinematograms, TV-static videos with a secret message, text readable only while playing but invisible in any paused frame, or the user asks what a ghost-font video says.
review
5-pass structured code review — correctness, security, performance, readability, consistency.
scaffold
Project-aware file generation. Reads existing codebase conventions (naming, structure, imports, exports, test patterns) then generates new files that match exactly. Wires generated files into the project's registration points.
design
Generates and maintains a design manifest for visual consistency. In existing projects, reads current styles and documents the design language. In new projects, asks a few questions and generates a starter manifest. The post-edit hook reads the manifest and flags deviations.
marshal
Meta-orchestrator that takes any direction — broad, specific, or vague — and autonomously chains skills and context into actionable work. Gathers context from codebase, docs, and memory. Only asks the user when it genuinely cannot proceed. Single-session orchestrator.
organize
Repository structure only: directory layout, file placement, naming conventions, and where-does-this-belong decisions. Detects the project's convention, audits files against it, and executes move plans with import-path updates. Never changes code inside files beyond the import updates a move forces; in-file…