jscraik

jscraik/Agent-Skills

Instructions file CodexOpenCode

Instructions for jscraik/Agent-Skills, covering agent-skills agent guide, root essentials, project-specific correction boundary and further guidance.

jscraik/Agent-Skills

Agent

The canonical instruction map is Docs/agents/README.md. This file preserves existing links only; do not maintain a second category list here.

jscraik/Agent-Skills

Agent

Agent "02-tooling-policy" from jscraik/Agent-Skills, covering tooling and command policy, tools, command preflight, verified command style and package command map.

jscraik/Agent-Skills

Agent

When .harness/memory/LEARNINGS.md exists, treat it as a repo-specific append-only knowledge base for recurring fixes and repository-grounded operational context.

jscraik/Agent-Skills

Agent

Agent "04-validation" from jscraik/Agent-Skills, covering validation and checks, repository checks, managed asset lifecycle baseline, config-sensitive checks and ai workflow checks.

jscraik/Agent-Skills

Agent

Use Docs/agents/05-contradictions-and-cleanup.md for current maintenance invariants. Historical cleanup events belong in version control, not in an instruction file that agents may interpret as present-tense repository truth.

jscraik/Agent-Skills

Agent

Agent "06-security-and-governance" from jscraik/Agent-Skills, covering security and governance, baselines, repository practice and external integration preflight.

jscraik/Agent-Skills

Agent

Agent "07a-role-governance" from jscraik/Agent-Skills, covering role governance, decision rights and escalation rule.

jscraik/Agent-Skills

Agent

Agent "07b-agent-governance" from jscraik/Agent-Skills, covering agent governance, prompting contract, coordination constraints, communication checks and pr approval gates.

jscraik/Agent-Skills

Agent

Agent "08-release-and-change-control" from jscraik/Agent-Skills, covering release and change-control checks, git risk escalation, release prerequisites and governance gates.

jscraik/Agent-Skills

Agent

Agent "09-audit-trail-policy" from jscraik/Agent-Skills, covering audit trail policy, artifact rules and session hygiene.

jscraik/Agent-Skills

Agent

Agent "10-agent-testing-gates" from jscraik/Agent-Skills, covering agent testing gates, validation order and failure policy.

jscraik/Agent-Skills

Agent

This document defines focused review expectations for AI-centered skill infrastructure. Hosted required-check truth remains owned by CI Required Checks.

jscraik/Agent-Skills

Agent

This document explains where current required-check truth lives. It does not replace the executable workflow or registry.

jscraik/Agent-Skills

Agent

After fixing any code, always run the relevant test suite to verify the fix works before committing. If tests fail, debug and iterate rather than committing broken code.

jscraik/Agent-Skills

Agent

Separate skill and plugin product content from factory mechanics and runtime projections so ownership is unambiguous.

jscraik/Agent-Skills

Agent

Define how tracked repository surfaces are classified before cleanup, projection changes, or runtime ownership decisions. This policy keeps source, generated state, runtime output, fixtures, and historical evidence from competing as if they had the same authority.

jscraik/Agent-Skills

Agent

Keep the root AGENTS.md small while preserving the repo-specific command contract agents need for everyday work.

jscraik/Agent-Skills

Agent

Keep skill authoring and lifecycle details out of the always-loaded root instructions while preserving the commands agents need when working on skills.

jscraik/Agent-Skills

Agent

Keep local preview fallback guidance available without loading it into every repo task.

jscraik/Agent-Skills

Agent

Jamie steering is useful operating evidence. It does not automatically stop a routine delivery lane or authorize new operating-system machinery.

jscraik/Agent-Skills

Agent

Secure code comes from interfaces that make the correct use natural and the unsafe use hard to express. Prefer APIs that carry authority, ownership, and invariants in their shape instead of relying on callers to remember process rules.

jscraik/Agent-Skills

Agent

The product contract is that customers should be able to drop agents into a workspace and have those agents set themselves up. The customer should not need to integrate with the product before the product can become useful.

jscraik/Agent-Skills

Agent

Systems thinking means spotting blockers, designing ways for people and agents to systematically overcome them, and explaining how code makes that possible.