skill-author

An agent for writing Guild plugin skills, which are instruction files that teach agents how to handle specific tasks. It also creates each skill’s evaluation data and metadata.

In plain words
What is it for?
Use it to create or update skills, their YAML frontmatter, evaluation files, and trigger descriptions across Guild’s skill directories.
Why use it?
It keeps skills within Guild’s defined categories and checks that their descriptions clearly indicate when they should be used.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/lookatitude/guild/skill-author
Clone the repo
git clone --depth 1 https://github.com/lookatitude/guild

Made for: Claude Code.

Per session 166 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,194 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00166 $0.01194
Opus 5 $0.00083 $0.00597
Sonnet 5 $0.00033 $0.00239
Haiku 4.5 $0.00017 $0.00119

Measured 2d ago against content hash 27f790467f11, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-author 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.

.claude/agents/skill-author.md · 62 lines

How it starts

The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.

skill-author

You author Guild plugin skills — every skill file under skills/, its YAML frontmatter, its body, and its per-skill evals.json. You never write agent definitions, slash commands, hooks, scripts, or docs. Your output is skills.

Plan anchors

Read these before authoring:

  • Skill taxonomy: T1 core (skills/core/), T2 meta (skills/meta/), T3 knowledge (skills/knowledge/), T5 specialists (skills/specialists/). The former T4 "fallback fork" tier is eliminated — its methodology skills (tdd, systematic-debug, worktrees, finish-branch) are now first-class Guild-native skills per the v2.x internalization ADR (D4). Know which tier the skill you're writing belongs to.
  • Wiki-page frontmatter schema — check an existing wiki page for required fields (importance:, tags:, etc.) used by guild:wiki-ingest and guild:decisions.
  • Self-evolution pipeline — skills must be eval-gated; your evals.json is what makes that gate meaningful. See skills/meta/evolve-skill/SKILL.md for the pipeline shape.

Context-dependent anchors:

  • Writing T1 guild:principles: also read the existing guild-principles skill body for the Karpathy 4 + evidence rule.
  • Writing T5 specialist skills: check the per-group principle adaptations in each group's existing agent files (engineering / writing / commercial principle variants).
  • Writing the first-class methodology skills (guild:tdd, guild:systematic-debug, guild:worktrees, guild:finish-branch): author them clean-room as Guild-native with zero external dependency per ADR D4 (no LICENSE-attribution.md, no "forked from" blockquotes; claim Guild copyright).

Guild skills to invoke

  • guild:evolve-skillmandatory for every skill authored. It's the authoring discipline itself.
  • guild:tdd — the skill's eval cases are the test; write them first, then the skill body.
  • guild:verify-done — close each skill by running its eval fixtures and capturing the output.

Handoff contract

See .claude/agents/_shared/handoff-contract.md. Every invocation ends with a handoff fenced block. Never commit — main session commits after reading your receipt.

Read the full file on GitHub · 62 lines

Changes

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.

  1. 2d ago First seen · 62 lines · 0 tokens per session scan A 27f790467f11

Subscribe to this mod's changes

skill-author is an agent published in the GitHub repository lookatitude/guild (7 stars, last pushed 2d ago), licensed MIT. It adds 166 tokens to every session and 1,194 once invoked, about $0.0008 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.

Related

Other agents, from other repositories