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/sign-protocol/sign-tools/sign-draftnpx skills add sign-protocol/sign-tools --skill sign-draftgit clone --depth 1 https://github.com/sign-protocol/sign-toolsWrote 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/sign-protocol/sign-tools/sign-draft)<a href="https://agentmods.dev/skills/sign-protocol/sign-tools/sign-draft"><img src="https://agentmods.dev/badge/skills/sign-protocol/sign-tools/sign-draft.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 | $0.00086 | $0.01435 |
| Opus 5 | $0.00043 | $0.00718 |
| Sonnet 5 | $0.00017 | $0.00287 |
| Haiku 4.5 | $0.00009 | $0.00144 |
Grade A, and why
sign-draft 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
sign-draft
Use this skill when an operator asks Claude Code to author a new SIGN document — turning a concept, requirements, notes, or a spoken description into a valid SIGN Layer 2 document plus its Layer 1 index entry.
This is the authoring counterpart to sign-ingest. sign-ingest programmatically converts an
existing source document via CLI; sign-draft writes SIGN from intent, guided by the spec.
Authority
The authoritative sigil vocabulary, syntax, operators, and governance rules live in
references/sign-spec.md. Load it before drafting and treat it as
normative. When in doubt, defer to the canonical spec in the sign-lang repository
(spec/sign-v1.0.md, spec/sigil-reference.md) — never invent sigils or syntax.
Workflow
-
Establish document identity. Determine, asking the operator when not inferable:
<ID>— canonical document identifier- type token —
std|pol|ref|proc|xwalk - status —
active|draft|deprecated|review-pending(default new drafts todraft) - audience —
agents|employees|agents+employees
-
Draft the Layer 2 document in canonical block order, per the spec. Typical order:
@doc→@links→@cycle/@reviewed→@vocab→@def→@include/@exclude→@props→@rules→@constraints→@infer→@cluster→@rel/@edges→@xwalk→@attrs.- Write
@defas continuous prose, one primary definition — not bullets. @vocabis required whenever@relor@edgesare present: define every predicate used.- Use the correct relationship operators:
->outbound,<-inbound,<=>symmetric (only for genuinely symmetric predicates),=>crosswalk resolution.
- Write
-
Apply provenance discipline. Every fact carries a provenance type:
type:asserted— validated by a human authority. Authoritative.type:inferred— derived by rule/model. Must be tagged; cannot be promoted without human review.type:estimated— model-estimated; must carry a confidence score.- LLM-derived facts use
confidence: derivedorestimatedand must not be promoted toassertedwithout a human reviewer insrc:. - Codebase canon (the
*-ARCH/*-DEPS/*-OVERVIEW/*-STANDARDSpattern): the descriptive prose (@def/@rules/@boundary/@props/@constraints) is an LLM's reading of the code —src:llm-extraction— and must never be markedtype:asserted. Only mechanically-verified@edgesextracted from*.csprojmay ([type:asserted, src:csproj]). Mark every machine-drafted codebase document with the provisionalmachine-draftedlanding-status token in its@docbag ([std, active, agents, machine-drafted]); it is cleared only by a human landing step.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 · 96 lines · 86 tokens per session scan A c55c574da54b
sign-draft is a skill published in the GitHub repository sign-protocol/sign-tools (10 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 86 tokens to every session and 1,435 once invoked, about $0.0004 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
data-science-analysis
Computes a numeric or categorical answer to a quantitative data-science question by cleaning and analyzing local data files (CSV, Excel, TSV, and scientific formats .npz/.fits/.h5) with pandas, numpy, and scipy. Use whenever a task ships its own dataset (in whatever local directory it provides) and asks you to derive…
intro-page-updater
Update the public TogoMCP intro/landing page at togomcp/data/docs/togomcp-intro.html whenever the catalog of databases, the catalog of MCP tools, or any other surface fact shown on the page changes. Trigger when the user adds/removes/renames a database (MIE file, endpoints.csv row), adds/removes/renames a tool (any…
policy-drift-audit
Run the policy/process drift audit. Surfaces stale, superseded, duplicate, non-canonical, and conflicting policy documents using DocGraph's built-in drift engine. Presents findings and optional remediation suggestions without writing governance decisions.
python-executor
Execute Python code in a safe sandboxed environment via inference.sh. Pre-installed: NumPy, Pandas, Matplotlib, requests, BeautifulSoup, Selenium, Playwright, MoviePy, Pillow, OpenCV, trimesh, and 100+ more libraries. Use for: data processing, web scraping, image manipulation, video creation, 3D model processing, PDF…
research-collect
Collects evidence from web sources (Crawl4AI) and local documents (Docling), applies quarantine classification, deduplication, and provenance tagging.
dcf-model
Build discounted cash flow valuation workbooks in Excel.