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/tale-project/tale/write-docsnpx skills add tale-project/tale --skill write-docsgit clone --depth 1 https://github.com/tale-project/taleWhat 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 | $0.00198 | $0.02009 |
| Opus 5 | $0.00099 | $0.01005 |
| Sonnet 5 | $0.00040 | $0.00402 |
| Haiku 4.5 | $0.00020 | $0.00201 |
Grade A, and why
write-docs 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 yesterday.
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 — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.
write-docs
A docs page exists to move a reader from an intent — "I want X" — to a verified outcome, and the docs read as one narrator: a calm, opinionated peer who shipped a similar product and is telling a capable stranger how this one works — not selling, not hand-holding, not spec-dumping. This skill owns the method for the source language; cross-locale authoring (translated labels, per-locale grammar) follows the repo's translation skill or conventions where present.
When this applies
Creating or editing any page in the repo's docs content tree, editing its navigation/sidebar config, adding or regenerating a docs screenshot or code example, reshaping a landing or overview page, or triaging a failing docs structural check.
Discover the repo's contract first
Before the first edit, locate — in this order — and read what exists:
- The docs content tree and its locale layout (which languages ship, which is the source).
- The docs guide nearest the content (
AGENTS.md,README,CONTRIBUTING) — it names the commands, the taxonomy, and the repo facts this skill deliberately does not carry. - The component registry the renderer actually supports — never author a tag the renderer doesn't register; an unregistered tag renders as broken text.
- The structural test suite that gates docs (and how to run it) — its rules are the floor.
- The product's message catalogs — UI labels in docs match the shipped strings character-for-character.
The repo's tests and guide are the contract; this skill's rules are the defaults that apply where the repo is silent.
Write a note first
Write a short planning note and record your answers to this form before you write the page:
- Journey & audience: Describe what the reader is trying to do when they land here, who they are, and which page type (playbook) fits.
- Verified claims: Describe which labels, defaults, routes, or limits you checked against the running product or its source — and what surprised you.
- Visuals & code plan: Describe which journey steps get a screenshot and which claims get runnable examples.
- Ripple: Describe which locales, nav entries, images, and sibling pages this touches.
What ships with it
5 files 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.
- yesterday First seen · 133 lines · 198 tokens per session scan A 3fb27502e263
write-docs is a skill published in the GitHub repository tale-project/tale (23 stars, last pushed yesterday), licensed MIT. It adds 198 tokens to every session and 2,009 once invoked, about $0.0010 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
audit-event
Use when adding or changing an audit event, the audit log schema, SIEM export, or anything writing to auditevents. Also use when adding a code path that grants or denies access, since denials are events. Triggers on "audit", "audit log", "auditevents", "SIEM", "journal", "access log", "who read what".
db-migration
Use when changing the PostgreSQL schema, adding a migration under packages/core/migrations, or touching row-level security, the grants table, tombstones, or the audit table. Triggers on "migration", "schema", "DDL", "RLS", "row level security", "ALTER TABLE", "new column", "index" in a database context.
mcp-tool
Use when adding or changing an MCP tool, the MCP transport, or MCP authorization in packages/mcp — tool schemas, tools/list, the Streamable HTTP endpoint, STDIO mode, OAuth, CIMD, EMA, or ID-JAG. Triggers on "MCP tool", "tools/list", "Streamable HTTP", "Mcp-Method", "CIMD", "DCR", "EMA", "ID-JAG", "resource server"…
open-core-boundary
Use when deciding whether a feature belongs in this open-source repository or in the private nacre-enterprise one, when adding an extension point, or when a change touches multi-tenancy, SSO, SCIM, document-level deny rules, EMA/ID-JAG, the audit log, the global admin, or quotas. Triggers on "enterprise"…
authz-change
Use when changing anything under packages/core/authz, the grants table, the effective-principals cache, or the vector pre-filter — the permission resolver, deny handling, scope inheritance, ACL tags, or the T1-T25 suite. Also use when a change elsewhere could affect who can see what, such as touching search filters…
config-var
Use when adding, renaming, or removing an environment variable, changing startup configuration validation, adding a Compose profile or service, or adding a Prometheus metric. Triggers on "environment variable", "NACRE", "config", "docker compose", "profile", "metrics", "healthcheck", "readiness".