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/boykush/scraps/ingestnpx skills add boykush/scraps --skill ingestgit clone --depth 1 https://github.com/boykush/scrapsWrote 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/boykush/scraps/ingest)<a href="https://agentmods.dev/skills/boykush/scraps/ingest"><img src="https://agentmods.dev/badge/skills/boykush/scraps/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 | $0.00062 | $0.01966 |
| Opus 5 | $0.00031 | $0.00983 |
| Sonnet 5 | $0.00012 | $0.00393 |
| Haiku 4.5 | $0.00006 | $0.00197 |
Grade A, and why
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 4d 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 — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ingest
Add a new scrap to the wiki and integrate it into the existing graph.
Implements Karpathy's Ingest primitive for Scraps: read a source, draft a new scrap, update related scraps with cross-links, and sanity-check the result. A single ingest typically touches the new scrap plus 1–5 existing scraps.
When to use
- "Write a scrap about X" / "Add a scrap on Y" (prompt source)
- "Summarize this article" / a pasted URL (web source)
- "Save this answer as a scrap" / file-back from a query (markdown source)
Source types
| Source | Provided as | First step |
|---|---|---|
| prompt | user's topic, instruction, or ready-to-use content (e.g., a /query synthesis) |
if the input is complete content, use as-is; otherwise gather context (search related scraps) and ask one clarifying question only when scope is unclear |
| URL | pasted link | WebFetch <url>, extract title and key content |
| term | an atomic term surfaced in discussion (often unknown until just now) | treat the term itself as the title; skip clarifying questions |
Workflow
-
Identify source and topic
- URL:
WebFetch→ use OGP / heading title as initial title - prompt: if the input is complete content (e.g., a
/querysynthesis), use as-is; if it's a topic, ask one clarifying question only when scope is unclear - term: the term is the title; do not ask clarifying questions (the discussion already established scope). Proceed to step 2 with the term as the search keyword.
- URL:
-
Research existing wiki state
scraps search "<keyword>" --jsonto find related scrapsscraps tag list --jsonto find relevant tags- Read 3–8 of the most related scraps via
scraps get "<title>" --json - Use
scraps get "<title>" --json headingsfirst when only the outline is needed
-
Decide title and ctx
- Pick a clear, atomic title
- If the title collides with an existing scrap, add a context folder:
scraps/<ctx>/<title>.md - ctx depth ≤ 3
-
Decide max-lines (familiarity heuristic)
- Count related scraps from step 2 (low: 0–5, medium: 6–15, high: 16+)
- Low → 5–7 lines (protect working memory)
- Medium → 10–12 lines (schema is forming, more detail welcome)
- High → 5–7 lines (avoid redundant explanation; prefer link-rich brevity)
- Skip this step if the user specified
--max-linesexplicitly (CLI flag is authoritative) - If the user signals depth in conversation ("go deeper" / "longer" / "厚めに" etc.), bump one band upward; "shorter" / "ライトに" etc. bumps downward. Conversation cue overrides the heuristic but stays below the explicit
--max-linesflag.
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.
- 4d ago First seen · 103 lines · 62 tokens per session scan A 05a8124b6c20
ingest is a skill published in the GitHub repository boykush/scraps (47 stars, last pushed today), licensed MIT. It adds 62 tokens to every session and 1,966 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 2026-08-30.
Other skills, from other repositories
processing-markdown
Processes Markdown files using mq, a jq-like query language for Markdown. Use when the user mentions Markdown processing, content extraction, document transformation, or mq queries.
web-scraping
Fetches web pages and extracts structured data using mq's toolchain (mq-crawl for fetching/crawling/JS rendering, mq for HTML-to-Markdown selector-based extraction, http() for in-query requests). Use when the user wants to scrape a URL, pull structured data out of a webpage, crawl a site, or turn HTML into…
okf
Author, maintain, and consume Open Knowledge Format (OKF) knowledge bundles — portable markdown + YAML frontmatter that both humans and agents read. Use when capturing project knowledge (services, APIs, schemas, metrics, runbooks, decisions) into an OKF bundle, when updating one after code or docs change, or when a…
shodh-memory
Persistent memory system for AI agents. Use this skill to remember context across conversations, recall relevant information, and build long-term knowledge. Activate when you need to store decisions, learnings, errors, or context that should persist beyond the current session.
validate
Check that an Open Knowledge Format (OKF) bundle is conformant with the v0.2 spec (§11). Use when asked to validate, lint, or check an OKF bundle, or before committing changes to one. Runs a deterministic Python checker — not an eyeball pass. Also migrates a v0.1 bundle to v0.2 in place with --migrate.
visualize
Render an Open Knowledge Format (OKF) bundle as a single self-contained, interactive HTML graph (viz.html) — concepts as nodes coloured by type and sized by body length, markdown links and bundle-internal sources as edges, a wiki-style detail panel with rendered markdown, v0.2 trust/lifecycle/provenance metadata, and…