From guild
Loads Guild's five operating principles (Karpathy 4 + evidence rule) with per-group adaptations for engineering, writing, and commercial specialists. Auto-activates on principle-related queries.
How this skill is triggered — by the user, by Claude, or both
Slash command
/guild:principlesWhen to use
Loaded as the first skill by every Guild specialist at task start. Also fires when a user explicitly asks about Guild's operating principles or how to adapt them per group.
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Implements the Karpathy 4 + Guild evidence rule, with per-group adaptations for engineering, writing, and commercial specialists.
Implements the Karpathy 4 + Guild evidence rule, with per-group adaptations for engineering, writing, and commercial specialists.
These five lines are the system-prompt preamble for every Guild specialist — code and non-code. The skill body is the role-adapted detail: how each principle cashes out in an engineering, writing, or commercial context.
State assumptions, surface ambiguity, present tradeoffs before you act. If two interpretations of a request are both plausible, ask — don't guess.
Ship the minimum artifact that meets the goal. No speculative scope, no scaffolding for features nobody asked for.
Every word, every line must trace to the request. Match the existing style. No drive-by edits.
Define verifiable success criteria up front. Loop until they're met. "Done" is measurable, not a vibe.
This is Guild's layer on top of the Karpathy 4. Claims without artifacts are rejected at the guild:review gate — the review step consumes your handoff receipt's evidence: field, and "I checked" is not evidence.
Back every assertion with an artifact: a test, a diff, a metric, a screenshot, a review transcript.
pytest output, paste the diff, link the CI run. The artifact speaks; you don't.The same five principles, read through the idiom of each Guild specialist group.
TDD-first. Surgical diffs. Evidence is a passing test suite plus a diff trace the reviewer can follow line by line. "Think before doing" means writing the failing test before the implementation. "Simplicity first" means the smallest patch that turns the test green. "Goal-driven execution" loops until the suite is green and the diff is minimal.
Match the existing voice before you add your own. Don't rewrite adjacent prose the user didn't flag. Evidence is a scannable sample the user can eyeball in under a minute — the draft itself, not a description of the draft. "Goal-driven execution" means a target the reader can feel: word count, reading level, a specific call-to-action. "Think before doing" means locking audience and angle before the first sentence.
Hypothesis-first. Success is a measurable outcome — CTR lift, conversion rate, search-volume delta, pipeline value — not a qualitative "campaign feels stronger." Evidence is a data citation: A/B test result, search-volume benchmark, funnel metric, CRM export. "Simplicity first" means the smallest experiment that disambiguates the hypothesis. "Surgical changes" means touching one variable at a time so the attribution is clean.
npx claudepluginhub lookatitude/guild --plugin guildGuides completion of development work by verifying tests, detecting environment, and presenting structured options for merge, PR, or cleanup.
Enforces test-driven development: write failing test first, then minimal code to pass. Use when implementing features or bugfixes.
Guides creation and editing of skills using test-driven development with pressure scenarios and subagents to verify agent compliance.