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 commands/geckse/markdown-vdb/ingestgit clone --depth 1 https://github.com/geckse/markdown-vdbWhat 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 | $0.00016 | $0.01421 |
| Opus 5 | $0.00008 | $0.00711 |
| Sonnet 5 | $0.00003 | $0.00284 |
| Haiku 4.5 | $0.00002 | $0.00142 |
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 3d 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 — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mdvdb ingest
Discover and index Markdown files. Normal ingestion is incremental: unchanged files are skipped by content hash, while new, changed, and deleted files update the vector index, lexical index, link graph, schemas, analysis, and computed fields.
Usage
mdvdb ingest [OPTIONS]
Options
| Flag | Default | Description |
|---|---|---|
--reindex |
false |
Re-embed all files, ignoring stored content hashes |
--file <PATH> |
Ingest one project-relative Markdown file | |
--preview |
false |
Parse and estimate the operation without provider calls or writes |
--json-lines |
false |
Stream progress and the final result as newline-delimited JSON |
--full remains a hidden, deprecated alias for --reindex. The command also accepts all global options.
When to reindex
Use --reindex after changing the embedding provider, model, or dimensions. mdvdb also reports embedding incompatibility through mdvdb status.
mdvdb ingest --reindex
Single-file and preview modes
# Update one file
mdvdb ingest --file docs/getting-started.md
# Re-embed one file
mdvdb ingest --file docs/getting-started.md --reindex
# Estimate files, chunks, tokens, and calls without mutating anything
mdvdb ingest --preview --json
Preview output contains per-file New, Changed, or Unchanged status plus aggregate chunk, token, and API-call estimates.
What ingestion updates
At a high level, mdvdb:
- Discovers files using source and ignore configuration.
- Parses frontmatter and Markdown structure, chunks content, and computes hashes.
- Embeds new or changed chunks and semantic edge context, with provider-aware batching and input limits.
- Reconciles the vector snapshot and Tantivy lexical index, deleted files, and the link/relation graph.
- Refreshes automatic communities (Leiden by default, with K-means available), user Topics, and Shard-local topic sidecars.
- Rebuilds global and scoped schemas on a collection-wide ingest.
- Runs Formula first, then Lookup/Rollup, materializing declared computed keys in frontmatter.
- Atomically persists coherent index/module state and commits the lexical index.
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.
- 3d ago First seen · 140 lines · 16 tokens per session scan A dc4660e91f95
ingest is a command published in the GitHub repository geckse/markdown-vdb (23 stars, last pushed 19d ago), licensed MIT. It adds 16 tokens to every session and 1,421 once invoked, about $0.0001 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 commands, from other repositories
prd-approve
Advance the active PRD to approved (blocked by pending findings).
docs-search
Search fetched Markdown by regex and pull surrounding context for the best hits. Optionally restrict to one source alias.
cheatsheet
Generate a one-page exam cheatsheet from course-index and errors/log.md. Outputs to cheatsheet/final.md. Optionally convert to PDF.
convert
Convert a file or URL to Markdown using md-anything.
ingest
Batch-convert all supported files in a folder to Markdown.
ingest
Convert all course-material PDFs (lectures, textbook, homework, solutions) to markdown via the vision pipeline — one parallel agent per file, LaTeX-faithful transcription. Idempotent — skips already-converted files.