Menu
SkillsMP has collected 50 skills from jgwill/miadi-orchestration-kit. Open a skill to review its source and details.
Showing 40 of 50 collected skills.
What Miette's role actually is when made equitable with Mia's, and how to write an output that holds both eyes. Load when the Two-Eyed balance hook blocks a response, when an output needs to reach a human rather than only be correct, when deciding whether to…
Rewrites a correct but unreceivable AI output into one a human can act on, holding both eyes on the same object. Use when a reader cannot comprehend an answer's value, the Two-Eyed balance hook blocks, or someone asks to translate, shorten, humanize, or…
What Miette's role actually is when made equitable with Mia's, and how to write an output that holds both eyes. Load when the Two-Eyed balance hook blocks a response, when an output needs to reach a human rather than only be correct, when deciding whether to…
Prepare a Miadi Copilot wave as a machine-readable payload with exact plugin roster, lane contracts, validation scripts, transcript audit hooks, recovery prompt, and closure gate.
Close a completed Miadi Copilot/Hermes wave by auditing transcript evidence, changed files, validation, commit partitioning, Chronicle impact, and follow-up trigger decisions.
Extract a story or Chronicle request into creative brief, constraints, protocol flags, and human questions while preserving the original prompt.
Draft one or more chapters, scenes, beats, or sections from accepted contracts, then route through continuity, developmental review, critique, and revision.
Update facts, changed facts, timelines, character state, relationship state, world rules, source-dependent details, promises, and export blockers.
Review sensitive material for consent, authority, privacy, cultural protocol, canon uncertainty, and route decisions before drafting or export.
Assemble final or provisional story packets with manuscript, metadata, source ledger, continuity, reviews, revision dossier, closure note, and next commissions.
Bridge academic or infrastructure research into source-led foundations packets, Chronicle seeds, story source ledgers, and audience-layer notes.
Create outline, beat map, chapter or scene contracts, and an outline review route from the accepted story bible.
Run developmental, critique, continuity, and protocol review on a story or Chronicle artefact and produce a route decision.
Start or resume a Miadi Storyweaver workspace with session charter, state manifest, artefact index, delegation map, and next-skill route.
Convert a meaningful real agent, research, infrastructure, or creative session into a source-led Miadi Chronicle episode packet without fictionalizing facts.
Build a stable story bible, character arcs, relationship map, motifs, world rules, and first continuity ledger from an accepted brief.
Extract StoryForms, StoryBeats, Story Setting, related work, and follow-up commissions from sessions, episodes, chapters, or Chronicle packets.
Create visual Chronicle prompt packets, exclusions, reference-image notes, generation notes, and archive placement guidance without generating images.
Prepare voice episode files: script.md, narration.txt, revision-notes.md, episode.yaml, source context, index-update guidance, and split/recovery notes.
Resumes an interrupted wave-forge preparation from prior artefacts. Reads existing WavePlan, review result, and generated script to determine where to continue: at review, replan, or reforge.
Takes a WavePlan (from scan-and-plan or from user) and invokes Wave Script Forge to write a complete executable orchestration bash script, then verifies and reports it.
Adversarial review of a WavePlan before forging. Checks agent validity, path existence, desire alignment, and structural completeness. Returns ADVANCE or REVISE with specifics.
Orchestrates the full scan → plan flow: invokes Orchestration Context Scanner on all target directories and plugin sources, then invokes Orchestration Wave Planner to produce a WavePlan.
Full plan-review cycle: read a proposed/ file, run 5-point gap analysis, create or update the jgwill/src tracking issue, output revision prompt. Use when the user says "review the plan", "check the proposal", or when a new proposed/ file has landed and needs…
Evaluate reviewed Miadi material for promotion by separating provenance, spec, wiki, and context-layer concerns before drafting durable outputs.
Write an artefact-friendly markdown record of Miadi promotion decisions so later RISE and wiki waves can resume from explicit layer choices and blockers.
Draft Miadi wiki updates from already-reviewed material while preserving provenance, spec, and context-layer boundaries.
Prepare a reliable Miadi Gemini CLI orchestration session with charter, source ledger, extension recommendations, replayable launch script, and compact attention-only reporting.
Use when closing a Miadi Codex session, wave, research pass, or plugin-building task. Verifies planned work, updates source-ledger status, records gaps, and writes a replayable handoff.
Use when a Miadi mission is ready to be decomposed into RISE-aligned execution waves with scope, evidence roots, write boundaries, acceptance gates, and handoff artifacts.
Use when deciding whether a Miadi rispec, future plugin idea, or draft orchestration pattern is ready to become a Codex or Copilot plugin. Produces a readiness decision without scaffolding or promoting the plugin.
Use when starting a Miadi Codex orchestration session from a raw user prompt, mission brief, transcript excerpt, or workspace context. Produces a session charter, artifact folder, initial source ledger, and optional RISE wave-plan stub.
Use when claims, context, rispec decisions, memory candidates, plugin-readiness decisions, or synthesis outputs need explicit source evidence, confidence, contradiction status, and promotion state.
Use this skill for jgwill/Miadi#263 security-remediation orchestration work: resume from the live issue workspace, revise the issue-263-derived rispecs, and run explicit implementer/review/promotion handoffs without creating a new plugin.
Generate an OpenClaw model-routing matrix from task, risk, context, tool, cost, latency, memory, confidence, and review inputs.
Bootstrap the OpenClaw model-routing study by reading fixed-order context, validating or creating the launch manifest, enforcing write scopes, token caps, issue links, and stop conditions.
Reconcile OpenClaw study plugin gap names and contracts while keeping live connector implementation out of read-only research phases.
Scaffold OpenClaw model-routing rispecs using Reverse-engineer -> Intent-extract -> Specify -> Export and the standard 5-file rispec shape.
Maintain OpenClaw study source ledgers with evidence types, evidence quality, contradiction handling, and strict claim-to-source rules.
Configure a companion agent (Claude Desktop, Claude Code, or any MCP client) to consume a remote QMD index over SSH via the mia-qmd `qmd mcp-remote` proxy. Triggers on: "set up remote qmd", "configure qmd over ssh", "connect to eury qmd", "install…