Templates
Copy-paste starting points with annotated fields. No email gate on the copy button, ever.
A copy-paste starting point for a well-formed agent skill: trigger-bearing description, procedure, boundaries, and verification.
A repository-root instruction file any agent runtime can read: setup, conventions, map, and constraints.
A Claude Code project instruction file: context, commands, style, and the accumulated gotchas that make sessions start smart.
A scoped Cursor rule: glob-targeted, rationale-bearing, with examples that show the rule applied.
A specification for a closed-loop workflow: steps, per-step verification, exit criteria, recovery, and budget.
A pre-adoption review checklist for any third-party skill: triggers, completeness, boundaries, evidence, docs.
The minimum governance layer for a production agent deployment: permissions, destructive-action policy, logging, human gates, data boundaries.
An AGENTS.md-skeleton tuned for OpenAI Codex CLI: same identity-layer content, Codex-specific execution notes.
A one-page specification of an agent memory system: types, storage, retrieval, decay, and privacy, decided before the first write.
A specification for a multi-stage agent workflow: actors, stages, skills per stage, handoffs, and verification gates.
The reporting skeleton for a reproducible agent benchmark: hypothesis, pinned method, results, limitations, raw data.
A self-audit of an agent system against the Agentic Maturity Model: failure inventory, loop and memory audit, top fixes.
The one-page manual for a production agent: mission, capabilities, boundaries, escalation, and ownership.
First-hour setup for a Claude Code project: CLAUDE.md, permissions, first skills, and hooks, in order.
The five-minute test for whether a skill actually fires: paraphrase matrix, fire-rate scoring, and description fixes.