Playbook As Code — Development skill for Claude Code
An open schema for contract negotiation playbooks, plus a clause locator, negotiation memo, redline-ready changeset and a Claude review skill.
How to install Playbook As Code
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open kevanwee/playbook-as-code and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Playbook As Code does
An open schema for contract negotiation playbooks, plus a clause locator, negotiation memo, redline-ready changeset and a Claude review skill. Three reference playbooks included.
Alternatives in Development
- Ic Memo — 投委会备忘录 — 8 章节结构化决策文档(含三情景回报 + Top 3 风险) 6.6k ★
- Commerce Agents — Reference blueprint for building shopping and merchant agents with Claude 2.6k ★
- Src Hunter Skill — 实战 SRC / 众测 / Bug bounty 漏洞挖掘 Claude Code skill — 19 个攻击类 playbook、305 个结构化 payload、263 个 WAF/EDR 绕过、2887 份 Ha 602 ★
README
playbook-as-code
Negotiation playbooks as data: an open YAML schema for "our position on each clause", a validator, a clause locator, and a Claude skill that reviews a contract against the playbook and emits a memo plus a redline-ready changeset.
- id: liability-cap
topic: Limitation of liability
severity: deal_breaker
ideal: Cap at 12 months' fees; exclusions only where required by law.
acceptable: Cap at 12 months' fees; carve-outs for confidentiality and IP indemnity; 2x super-cap for data.
walk_away: Uncapped, or data-protection breach carved out with no super-cap.
detect:
any_of: ["aggregate liability", "shall not exceed", "limitation of liability"]
fallback_language: |
...each party's total aggregate liability ... shall not exceed the Fees paid or payable
in the twelve (12) months immediately preceding the event giving rise to the claim.
authority: General Counsel
jurisdiction_notes:
SG: UCTA s 11 reasonableness applies; an unreasonably low cap risks being struck out.
Why
Every legal team has a playbook. It lives in a Word table nobody opens, so every review starts from memory. The commercial products that solve this (Luminance, Robin, Spellbook) keep the playbook format proprietary. There is no open schema, which means no portability, no diffing your positions against a counterparty's, and no way for an AI assistant to review against your actual standards rather than generic ones.
The schema is the contribution. The tooling around it is deliberately thin.
What is in the box
| Piece | What it does |
|---|---|
schema/playbook.schema.json |
JSON Schema, generated from the pydantic model. Validate in any language. |
playbooks/ |
Three real playbooks: SaaS vendor-side, SaaS customer-side, mutual NDA. Singapore-law notes. |
playbook validate |
Fails loudly on duplicate ids, empty detectors, bad severities. |
playbook scaffold |
Splits a contract into numbered clauses, finds which cla |
Related Skills
Rogue Contract Review Risk Scoring Skill
Review contracts with clause-by-clause risk scoring, market benchmarks, negotiability ratings, and redline sug
Superdialog Playbook Skill
Claude Code / Codex skill for authoring and reviewing SuperDialog voice-agent playbook YAMLs — schema referenc
Word Choice Skill
Modal verb taxonomy — applies to every contract review and redline analysis:
Rogue Legal Work Plugin Skill
In-house legal workflows for contract review against playbooks, NDA triage with GREEN/YELLOW/RED ratings, comp
Arabic Letter
Draft a formal Arabic document (letter, CV, report, memo, minutes, or simple contract) using the arabic-formal
Monday
Write the weekly memo. Week-over-week Search Console (and GA4 when configured), the status of every edit Caddi
Related Agents
Clause
Contract clause analysis — redlining, risk scoring, negotiation playbooks
Uft Converter
Assembles the conversion plan for a UFT test from the three upstream artifacts and writes 04-proposal.md plus
Code Locator
Read-only sweep that maps where things live in a codebase — "where is X defined", "what calls Y", "every refer