joycraft

Invoked by gather-context or the human after a knowledge gap surfaces — author one long-form reference doc and wire a pointer into AGENTS.md's Context Map

Invoked by session-end or the human after a fact surfaces — route it to the correct context document (production map, dangerous assumptions, decision log, institutional knowledge, troubleshooting)

Invoked at the design bookend by decompose's decision gate or the human directly — turn open questions into a decision dossier; every decision terminates clarified, backlogged, or discarded

Break a feature brief into atomic specs — small, testable, independently executable units

Design discussion before decomposition — produce a ~200-line design artifact for human review, catching wrong assumptions before they propagate into specs

Invoked by tune, optimize, or session-end to convert eligible boundary prose into machine-checked deny patterns — not a user entry point.

Invoked by tune's roadmap or the human opting into Level 5 — set up the autofix loop, holdout scenario testing, and scenario evolution from specs

Invoked before a constrained implementation session or by the human directly — generate NEVER rules and deny patterns to prevent agent overreach

Guided feature development — interview the user, produce a Feature Brief, then decompose into atomic specs

Invoked by tune's roadmap or the human directly — semantic self-audit of harness overhead; per control it assigns a disposition (KEEP/ONE_HOME/LOAD_LATER/MAKE_A_CHECK/PROBATION/RETIRE) and an evidence label, advisory only

Invoked by design/decompose or the human directly — produce objective codebase research by isolating question generation from fact-gathering

Assess and upgrade your project's AI development harness — score 7 dimensions, apply fixes, show a harness maturity roadmap

Invoked by session-end or the human directly — spawn an independent verifier subagent to check an implementation against its spec, read-only

Invoked by gather-context or the human after a knowledge gap surfaces — author one long-form reference doc and wire a pointer into CLAUDE.md's Context Map

Invoked by session-end or the human after a fact surfaces — route it to the correct context document (production map, dangerous assumptions, decision log, institutional knowledge, troubleshooting)

Invoked at the design bookend by decompose's decision gate or the human directly — turn open questions into a decision dossier; every decision terminates clarified, backlogged, or discarded

Break a feature brief into atomic specs — small, testable, independently executable units

Design discussion before decomposition — produce a ~200-line design artifact for human review, catching wrong assumptions before they propagate into specs

Invoked by tune, optimize, or session-end to convert eligible boundary prose into machine-checked deny patterns — not a user entry point.

Invoked by tune's roadmap or the human opting into Level 5 — set up the autofix loop, holdout scenario testing, and scenario evolution from specs

Guided feature development — interview the user, produce a Feature Brief, then decompose into atomic specs

Invoked by tune's roadmap or the human directly — semantic self-audit of harness overhead; per control it assigns a disposition (KEEP/ONE_HOME/LOAD_LATER/MAKE_A_CHECK/PROBATION/RETIRE) and an evidence label, advisory only

Invoked by design/decompose or the human directly — produce objective codebase research by isolating question generation from fact-gathering

Wrap up a session — capture discoveries, verify, prepare for PR or next session

Assess and upgrade your project's AI development harness — score 7 dimensions, apply fixes, show a harness maturity roadmap

Invoked by session-end or the human directly — spawn an independent verifier subagent to check an implementation against its spec, read-only

Invoked by gather-context or the human after a knowledge gap surfaces — author one long-form reference doc and wire a pointer into AGENTS.md's Context Map

Invoked by session-end or the human after a fact surfaces — route it to the correct context document (production map, dangerous assumptions, decision log, institutional knowledge, troubleshooting)

Invoked at the design bookend by decompose's decision gate or the human directly — turn open questions into a decision dossier; every decision terminates clarified, backlogged, or discarded

Break a feature brief into atomic specs — small, testable, independently executable units

Design discussion before decomposition — produce a ~200-line design artifact for human review, catching wrong assumptions before they propagate into specs

Invoked by tune's roadmap or the human opting into Level 5 — set up the autofix loop, holdout scenario testing, and scenario evolution from specs

Invoked before a constrained implementation session or by the human directly — generate NEVER rules and deny patterns to prevent agent overreach

Guided feature development — interview the user, produce a Feature Brief, then decompose into atomic specs

Invoked by tune's roadmap or the human directly — semantic self-audit of harness overhead; per control it assigns a disposition (KEEP/ONE_HOME/LOAD_LATER/MAKE_A_CHECK/PROBATION/RETIRE) and an evidence label, advisory only

Invoked by design/decompose or the human directly — produce objective codebase research by isolating question generation from fact-gathering

Assess and upgrade your project's AI development harness — score 7 dimensions, apply fixes, show a harness maturity roadmap

Invoked by session-end or the human directly — spawn an independent verifier subagent to check an implementation against its spec, read-only

Structured bug fix workflow — triage, diagnose, discuss with user, write a focused spec, hand off for implementation

Set up Joycraft for a team — scaffold per-area folders, owner conventions, and a team-facing CONTRIBUTING doc. Run once when adopting Joycraft on a multi-dev project.