MrMaxie
- 20 skills
- 0 followers
- 9 hours ago last updated
- ▌ Adopt Arcantry · mrmaxie bundleInspect, adopt, validate, or remove Arcantry at shared, private, or configuration-free project scope while preserving repository-owned files.
- ▌ Agent Self Improve · mrmaxie bundleDiagnose recurring communication failures in supplied or available agent conversations and stage minimal guidance corrections for approval. Use only when explicitly requested.
- ▌ Design Terminal UX · mrmaxie bundleDesign or audit terminal interactions when navigation, status, errors, recovery, accessibility, destructive actions, or small-screen behavior affect usability.
- ▌ Select Task Skills · mrmaxie bundleChoose at most three available skills whose distinct workflows materially help a complex, multi-domain, or unfamiliar task. Do not use for simple work or install anything.
- ▌ Design Coherent Gui · mrmaxie bundleDesign, implement, or audit graphical interfaces when information density, hierarchy, spatial composition, control sizing, or component consistency affects usability. Do not use for CLI or TUI work or visual styling alone.
- ▌ Capture Project Work · mrmaxie bundleRoute one project thought to todo, OpenSpec, changelog, or skill improvement at its current commitment level without duplication or premature promotion.
- ▌ Audit Skill Portfolio · mrmaxie bundleInventory an installed skill portfolio, distinguish aliases from identity conflicts, and recommend evidence-backed routing or consolidation actions. Use only when explicitly requested and remain read-only.
- ▌ Evaluate Skill Change · mrmaxie bundleMeasure whether a candidate skill improves routing and behavior over its baseline before adoption. Use only when explicitly requested and never install the candidate automatically.
- ▌ Write Product Content · mrmaxie bundleWrite or revise reader-facing product documentation, release text, README copy, UI copy, or technical explanations using concrete evidence and no internal process narration.
- ▌ Maintain Release Story · mrmaxie bundleAlign accepted OpenSpec outcomes, unit-aware SemVer, distributable versions, and consumer-facing changelog prose when planning, preparing, or auditing a release without publishing it.
- ▌ Protect Local Boundary · mrmaxie bundleProtect a repository's `.local/` privacy boundary whenever work may create, read, write, copy, move, derive, watch, build, test, package, document, or promote `.local` content. Use for every planned or actual operation involving `.local`; do not use for a conceptual mention with no operation.
- ▌ Capture Repeatable Work · mrmaxie bundleExtract repeated workflow friction from supplied or available conversations and propose the smallest durable carrier without implementing it. Use only when explicitly requested.
- ▌ Maintain Agent Guidance · mrmaxie bundleAudit and update universal AGENTS.md guidance at the narrowest user, repository, nested, or private scope through exact approval-gated proposals.
- ▌ Promote Todo To Openspec · mrmaxie bundleGroup and transform selected todo.txt items into approval-gated OpenSpec changes while preserving privacy and avoiding one-to-one promotion assumptions.
- ▌ Audience Scope Discipline · mrmaxie bundleProtect the audience, scope, privacy boundary, and established structure of product-facing or derived artifacts when implementation or content could drift.
- ▌ Reconcile Project Sources · mrmaxie bundleExplain drift between shared and private OpenSpec, changelog, or todo sources and prepare one explicit transition that preserves both until apply.
- ▌ Productize Repeatable Work · mrmaxie bundleImplement one approved repeatable-work candidate in its chosen durable carrier without reopening its scope. Use only with an accepted stable candidate id.
- ▌ Stage Code Review Findings · mrmaxie bundleHold evidence-backed code-review findings in chat until the user selects what may become an external draft. Use for local, draft, or pre-publication review.
- ▌ Verify Work Proportionally · mrmaxie bundleMatch verification layers to impact, uncertainty, reversibility, and audience when tests alone cannot prove the requested outcome or acceptance boundary.
- ▌ Forge Skill From Conversations · mrmaxie bundleTurn one recurring workflow from supplied or available conversations into a focused, evaluated skill candidate. Use only when explicitly requested and require approval before creation or adoption.