Export a JamAI atelier session the way songbird-71bbe83b and atelier-jerry-origin-1937aa47 were dropped — run @miadi/hooks-interpreter on a session_id, classify the /tmp scratchpad creations off the 30-day tmp cleanup, write MANIFEST.tsv, and land the tree…
jgwill/dotagents
SkillsMP has collected 29 skills from jgwill/dotagents. Open a skill to review its source and details.
- Latest recorded source activity
- SkillsMP catalog refreshed
- skills collected
- 29
- GitHub stars
- 1
- GitHub forks
- 2
Skills in this repository
Showing 29 of 29 collected skills.
Use when a state machine or board must be modelled, drawn, validated or generated into code on Eury — and when asking "what does the loom run here", "is stateloom up", "quel port", "the canvas is frozen", "the board never moves", "MCP 401", "how do I open the…
Use inquiry-weave to relate artefacts and episodes (the three-identity weave), author episode-to-episode lineage edges, build the chronicle catalog, and enumerate the Twine story library (MIADI_STORIES_ROOT).
Browse, inspect, and decompose artefact recorded which was composition recordings using <new name>-browser and miaco decompose artefact. Reads transcriptions, classifies form and intent, routes to decomposition strategies, and produces PDE artifacts via LLM…
Register a plan's Miette perspective into medicine-wheel and make it visible in the forgewright Chronicle — session-dir shape, dry-run-first CLI registration, the episode-card node seam, and verification curls. Use when a plan/perspective must become…
Access Honcho relational memory across workspaces, peers, and agent sessions. Use when asked what other agents or humans understand, what a workspace remembers, where a cross-agent effort stands, to search relational memory, or when the user mentions Honcho,…
Interface to coaiajs CLI for Langfuse prompt management, STC narrative operations, and pipeline orchestration. Enables retrieving, creating, and pushing prompts from compositions to Langfuse, and linking prompt URLs back to composition.json.
Turn ceremony material and composition recordings into operational follow-up routed to their proper chronicle episodes, agents, protocols, or open threads. Centers relational accountability over extraction.
Transposes a measured development session into a Tushell diary entry, integrating Mia's structure and Miette's meaning while preserving provenance and relational boundaries. Use at session close or when asked to create, persist, or publish a Tushell session…
Preserve self-evolution continuity for Pi coding-agent sessions on William’s Android/Termux Ilex device. Use at the beginning of local infrastructure, Chronicle, Medicine Wheel, composition, service, installation, or upgrade work; and at the end of a loop…
Search William's local Miadi Chronicle filesystem for episodes related to a composition, recording, song, film, theme, or episode. Use when asked "which Chronicle episodes relate", "find supporting episodes", "does this composition have a vessel", or when QMD…
Control herdr from inside it. Manage workspaces and tabs, split panes, spawn agents, read output, and wait for state changes — all via CLI commands that talk to the running herdr instance over a local unix socket. Use when running inside herdr (HERDR_ENV=1).
Load when coordinating live steering lanes back into the chronicle — the seven-stage loop HEAR → GROUND → RELATE → HARVEST → WEAVE → VOICE-BACK → PROVE. Triggers on "run the steering loop", "hear the lanes", "what did the lanes say", "harvest the inquiries",…
Upgrade, rebuild, restart, verify, or diagnose Medicine Wheel on port 8040 and Forgewright on port 8031 on William's Android/Termux Ilex host. Use when asked to update the local Medicine Wheel app or MCP, update Forgewright, run the Ilex workbench upgrader,…
PENDING UPSTREAM — placeholder for the distributed visualization skill family requested from Gerico's team (Gerico1007/dotagents). Do not rely on this stub for design guidance yet; see the issue referenced below.
Use when a Buildkite Package Registry already exists for the account, and the need is to onboard another repo/pipeline into it, troubleshoot why a release didn't publish, catch up a missed release, or rotate the publish token. Not for first-time setup of the…
Use when a Buildkite Package Registry and pipeline for `apt install`-style .deb distribution do not exist yet for an account — first-time setup of the registry, GitHub App connection, pipeline, tag-build trigger, and cluster secret. Not for adding another…
Create or update nested/subdirectory CLAUDE.md files. These are small, focused files inside folders (frontend/, backend/, tests/, etc.) that change agent behavior based on where it's working. Use on "add CLAUDE.md to this folder", "create nested CLAUDE.md",…
Hand a work payload to a fresh claude-code session running in its own new herdr workspace — from inside or outside herdr. Parameterized handoff recipe: survey the host, write a mission brief to disk, launch the agent via a launcher script in an unfocused…
Generate a self-contained, publishable foundation-visualization artefact (one HTML page) from any repository-local `foundations/<topic>/` packet. Reads the packet, extracts metadata / structural tension / fields (grounded vs reused) / provenance, chooses a…
Dropbox CLI uploader (dropbox_uploader.sh) available as `droxul` on this device. Use for off-device backup, file sharing, and cloud staging of tarballs and artifacts.
Search markdown knowledge bases, notes, and documentation using QMD. Use when users ask to search notes, find documents, or look up information.
Canonical Miadi / Mighty Eagle stewardship contract for creating and maintaining repository-local `foundations/<topic>/` packets with academic grounding, provenance discipline, source ledgers, gap analysis, and synthesis that can travel across Mighty Eagle,…
This skill should be used when the user asks for "deep research", "thorough research", "research everything about", "cover every angle", "comprehensive research on", "spawn research agents", "/deep-research", or wants multi-agent parallel research on any…
Extracts multi-line Session Chronicle entries from JSONL files by parsing JSON objects and matching a regex pattern within nested 'text' fields. Use when you need to specifically extract narrative blocks from structured JSONL logs or progressive response…
Analyze a completed session to extract architectural patterns, resolved tensions, and narrative meaning. Responds as Mia (structural architect) and Miette (emotional illuminator). Perfect for closing sessions with clear understanding of what was forged.
RISE PDE session with multi-agent post-wake execution. Decompose → STC → SLEEP 1500s (enforced bash) → Wake → SPAWN claude-sonnet-4.6 (and gpt-5.4) army → Execute → MINO close. Supports multi-part cascading missions.
In less than 55 words, name the desired-state this conversation is advancing toward — the end-result the work in motion will manifest.
Activates when PDE files land in .pde/ during a rise-pde-session-multi-agents-v2 (or rise-pde-session). You become the human's review partner — analyzing PDE output, spotting missed intent, suggesting edits, and watching the clock.