HelixDevelopment
OrganizationConstitution, AGENTS.MD and CLAUDE.MD universal agent rules and constraints
Categories
Indexed Skills (3)
action-prefix-system
Use when a prompt begins with an UPPERCASE action token (BACKGROUND / REMINDER / ISSUE / BUG / TASK / a sub-system name) in any of the six §11.4.140 forms — `NAME :: rest`, `PREFIX::NAME :: rest`, `/NAME rest`, `/PREFIX::NAME rest`, `NAME ---> rest`, or the single-colon `NAME: rest` — or when the operator asks how the action-directive / slash-command system works, how to add a new directive, why a bare `/name` did not resolve, why an unregistered `TODO:` line correctly did nothing, or what a grammar-shaped token that is NOT registered should do. Also use before inventing ANY expansion for an unknown UPPERCASE token.
reporting-workable-items
Use whenever anything trackable is reported or discovered — a bug, a defect, a regression, a feature request, an internal task, a gap, "this is broken", "we should add X", "remember to do Y" — or when the operator invokes /issue, /bug, /task (or their /helix: and /default- forms, or the ISSUE/BUG/TASK action prefixes). Use it BEFORE answering a report in prose, and use it to check whether a report you are about to acknowledge has actually been turned into a real tracked item.
workable-item-lifecycle
Use when moving a tracked item through its lifecycle — starting work, marking it ready for testing, closing it, reopening it, marking it obsolete or operator-blocked — or when asked which Status/Type value is legal, how to close a Feature vs a Bug, why the Issues/Fixed docs must be regenerated rather than hand-edited, what a reopen must record, or what belongs in a per-item testing diary.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.