Skip to content

Comments

feat: add writing-skills skill, clarify subagent dependency#4

Open
coctostan wants to merge 2 commits intomainfrom
feat/testing-infrastructure
Open

feat: add writing-skills skill, clarify subagent dependency#4
coctostan wants to merge 2 commits intomainfrom
feat/testing-infrastructure

Conversation

@coctostan
Copy link
Owner

Summary

Incorporates the writing-skills skill from obra/superpowers — TDD applied to process documentation.

Changes

New: writing-skills skill (7 files)

  • SKILL.md — Main skill: TDD mapping, CSO guidance, bulletproofing against rationalization
  • anthropic-best-practices.md — Anthropic's official skill authoring guide
  • persuasion-principles.md — Psychology research for designing effective skills
  • testing-skills-with-subagents.md — Pressure testing methodology
  • graphviz-conventions.dot — Graphviz style guide for flowcharts
  • render-graphs.js — Script to render flowcharts to SVG
  • examples/CLAUDE_MD_TESTING.md — Worked example of testing doc variants

Adapted from upstream:

  • superpowers:name/skill:name cross-references
  • TodoWriteplan_tracker
  • Agent-neutral language (no Claude Code-specific refs)

README improvements

  • New Optional: Subagent Tool section right after Install — clearly documents that pi-superpowers does not include a subagent tool
  • 🤖 markers in the skills table show which 4 skills use subagents
  • Points users to pi's example subagent extension or manual dispatch
  • Removed buried/duplicate subagent section

Version bump

0.1.1 → 0.2.0

Tests

All 53 tests pass (34 plan-tracker + 19 skill validation).

Incorporates the writing-skills skill (TDD applied to process documentation)
adapted from https://github.com/obra/superpowers/tree/main/skills/writing-skills

Files:
- SKILL.md - Main skill with TDD mapping, CSO, bulletproofing guidance
- anthropic-best-practices.md - Anthropic's official skill authoring guide
- persuasion-principles.md - Psychology research for skill design
- testing-skills-with-subagents.md - Pressure testing methodology
- graphviz-conventions.dot - Graphviz style guide for flowcharts
- render-graphs.js - Script to render flowcharts to SVG
- examples/CLAUDE_MD_TESTING.md - Worked example of testing doc variants

Adaptations from upstream:
- superpowers:name -> /skill:name cross-references
- TodoWrite -> plan_tracker
- Agent-neutral language (no Claude Code-specific refs)

Also updates README.md:
- Added writing-skills to skills table
- Updated Subagent Dispatch section to document that subagent tools
  are not included; points to pi's example extension or manual dispatch
- Move subagent info to prominent 'Optional: Subagent Tool' section
  right after Install, before the skills table
- Mark subagent-dependent skills with 🤖 in the table
- Explain that skills work as process guides without a subagent tool
- Remove duplicate subagent section from lower in README
- Bump version 0.1.1 -> 0.2.0 (new writing-skills skill)
@socket-security
Copy link

Review the following changes in direct dependencies. Learn more about Socket for GitHub.

Diff Package Supply Chain
Security
Vulnerability Quality Maintenance License
Updated@​mariozechner/​pi-coding-agent@​0.52.9 ⏵ 0.53.073100100 +196100
Updated@​mariozechner/​pi-ai@​0.52.9 ⏵ 0.53.096100100 +196100
Updated@​mariozechner/​pi-tui@​0.52.9 ⏵ 0.53.097100100 +196100

View full report

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant