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/smixs/autograph/researchgit clone --depth 1 https://github.com/smixs/autographWhat 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 | $0.00047 | $0.01426 |
| Opus 5 | $0.00023 | $0.00713 |
| Sonnet 5 | $0.00009 | $0.00285 |
| Haiku 4.5 | $0.00005 | $0.00143 |
Grade A, and why
research 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 — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/autograph:research — interactive bootstrap
You are orchestrating an interactive bootstrap of a chaotic or empty Obsidian vault. The user invoked /autograph:research $ARGUMENTS. Treat $ARGUMENTS as the vault path. If empty, ask for it with AskUserQuestion.
This command only runs when the vault lacks a usable schema and structure. For structured vaults, redirect the user to the normal bootstrap workflow (references/bootstrap-workflow.md in the autograph skill).
Workflow
Phase 1 — Gate (is /research the right tool?)
Run discovery to see what we have:
uv run <skill-dir>/scripts/discover.py "$VAULT" --verbose > /tmp/autograph-discovery.json
uv run <skill-dir>/scripts/research.py plan "$VAULT" > /tmp/autograph-plan.json
Read /tmp/autograph-plan.json. It contains gate with fields:
files_total,folders_totalhas_schema_json(bool)frontmatter_coverage(0–1)verdict:"chaos"|"structured"|"empty"
If verdict == "structured": stop. Tell the user:
This vault already has a workable structure (N files, X% with frontmatter, schema.json present).
/researchis for empty or chaotic vaults. Usediscover.py→generate_schema.py→swarm_prepare.pyfrom the autograph skill instead.
Do not proceed to Phase 2 unless the user overrides.
If verdict == "empty" (≤ ~20 files): skip Phase 3 swarm — go straight to Q&A (Phase 2) + manual schema draft.
If verdict == "chaos": continue to Phase 2.
Phase 2 — First-wave Q&A
Use AskUserQuestion to establish intent. Ask up to 4 at once:
- Purpose — personal / work / research / mixed
- Expected domains — free-form list, show 3–5 folder names already present as hints
- Content language — ru / en / mixed / other
- Schema ambition — minimal (3–5 types) / standard (6–10) / full taxonomy (10–15)
Store answers in memory — you will feed them to Phase 4.
Phase 3 — Swarm exploration (only for chaos)
/tmp/autograph-plan.json contains a manifests array. Each manifest has a non-overlapping subset of files (≤ 20 each) bin-packed by folder. Spawn parallel Agent calls — one per manifest, up to 5 in a single message — with subagent_type: "general-purpose".
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 · 111 lines · 47 tokens per session scan A 50e02515fa86
research is a command published in the GitHub repository smixs/autograph (67 stars, last pushed 28d ago), licensed MIT. It adds 47 tokens to every session and 1,426 once invoked, about $0.0002 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
slm-loop
Run a task as a bounded, gate-verified loop backed by SuperLocalMemory — iterate until an independent gate passes, never on the agent's own claim.
gsd:graphify
Build, query, and inspect the project knowledge graph in .planning/graphs/.
gsd:discuss-phase
Gather phase context through adaptive questioning before planning.
gsd:config
Configure GSD settings — workflow toggles, advanced knobs, integrations, and model profile.
gsd:autonomous
Run all remaining phases autonomously — discuss→plan→execute per phase.
gsd:capture
Capture ideas, tasks, notes, and seeds to their destination.