kalepail

Change the golden Q→A corpus without codifying lies: classify the truth domain, triangulate across independent source classes, encode disputed or unverifiable facts honestly, and land provenance-bearing edits to the owned per-case corpus. Use when editing eval golden answers, key facts, avoid clauses, sources, grader notes, or truth metadata.

Resolve a "Live service drift detected" issue — the daily CI that diffs the committed catalog against the live upstream surfaces (Lumenloop, Stellar Light/Scout, Stellar Docs). Regenerate the artifacts, classify the drift, decide whether it needs a policy or routing-baseline change or is a clean mechanical bump, dual-verify, commit, deploy when authorized, verify production, and close. Use when a drift issue is filed, when refreshing inventory, or when deciding if an upstream version bump is routing-neutral.

Run a full eval round on stellar-raven-codemode from Codex, Claude Code, or another CLI agent — pick the right instruments (routing gate, QA headline, agentic, plan, live-data), distinguish the orchestrating agent from the spawned answering and judge agents, record results, review every answer/verdict, triage failures to root cause, and file evidence-backed upstream service-improvement findings in improvements/. Use when asked to run evals, check gates, measure a scoring/catalog/executor change, review QA verdicts, understand eval model roles, or close an eval round. The primary artifact of every round is upstream findings, not the scores.

Coordinate a full truth-maintenance pass for stellar-raven-codemode across live drift, evals, golden questions/answers, and improvements/issues/PR follow-up. Use when asked to ensure repo truth is current, review the current state of upstream findings or PRs the repo opened, run a multi-lane maintenance round, or orchestrate CI-like review work with Solo agents instead of one monolithic context window.

Maintain the improvements/ upstream-findings pipeline for stellar-raven-codemode. Use when filing or updating service-improvement findings, resolving intake targets, maintaining improvements/intake.json, interpreting improvements lint failures, running recurrence probes, regenerating improvements/INDEX.md, reviewing upstream issues or PRs opened from findings, or handling eval-round and drift-refresh improvement maintenance.

Scouts the Stellar ecosystem before you build. Validates ideas against existing projects, matches open SCF-funded RFPs, drafts SCF pitches, finds audit firms, compares RFPs, recommends Soroban SDK skills, and cites primary sources (audits, SEPs, papers). Use whenever someone asks what to build on Stellar, vets an idea, preps an SCF grant application or hackathon entry, needs an auditor, or needs prior-art / security findings on a Soroban protocol.

Investigate live Cloudflare Workers observability for stellar-raven-codemode. Use when reviewing production MCP request logs, playground/demo telemetry, traces, Ray IDs, request headers, telemetry query results, eval observability, agent-run forensics, or whether Cloudflare metadata is enough without app-level correlation IDs.

Connect your AI assistant (Claude, ChatGPT, Gemini, Cursor, or any MCP client) to LumenLoop's free read-only Stellar ecosystem MCP at https://mcp.lumenloop.com and learn its 18 query tools. Use when the user wants to set up or troubleshoot the lumenloop connector, asks what Stellar ecosystem data is available, or needs to know which lumenloop skill fits a task.

Route broad or mixed Solo and SoloTerm requests across focused project, process, agent, durable-state, and automation skills. Use when choosing among Solo skills, deciding between one agent and an orchestrated cohort, learning Solo, or handling several Solo capabilities in one task. Do not use for non-Solo work, or when one clearly scoped Solo action already maps to a single focused skill.

Conduct and synthesize deep, exhaustive, comparative, or cross-checked research into a decision or contested question using independent evidence lanes, claim-level citations, contradiction resolution, and stress-tested conclusions. Use when the deliverable is a cited research synthesis or recommendation, including deep multi-source Stellar-ecosystem investigations. Do not use for a narrow lookup or single-source read — one fact, version, API signature, or docs page, even a Stellar or Soroban one; answer those directly from the authoritative source. Do not use for implementation, code review, or generic multi-agent coordination.

Plan and coordinate multi-agent work through Solo with bounded workers, durable state, locks, timers, evidence review, and handoffs. Use for parallel research, disjoint implementation lanes, cross-model review, cross-project cohorts, or delegation through several Solo agents. Do not use solely because research is deep or cited—citation and synthesis discipline come from a research method, not from orchestration—nor for single-agent, sequential, or shared-file work.

Configure, health-check, launch, prompt, inspect, resume, summarize, follow up with, and close one Solo-managed agent session. Use for built-in or custom CLI agent setup, launchability, one-off flags, input, output, status, summaries, idle or blocked handling, and owned-agent cleanup. Do not use for multi-agent orchestration or non-Solo agents.

Use agent-browser with Chrome DevTools Protocol virtual WebAuthn authenticators to test passkey, WebAuthn, Stellar smart account, wallet creation, wallet restore, and browser signing flows. Trigger when a task mentions passkeys, WebAuthn, virtual authenticators, Stellar Smart Account Kit browser tests, or agent-browser passkey automation. Do not use for browser flows without a passkey or WebAuthn ceremony—password logins and extension-wallet connects route to agent-browser alone.

Operate Solo-managed commands and terminals through start, stop, restart, rename, spawn, input, bulk actions, trust gates, auto-start, crash restart, and file-watch restart. Use when launching a dev stack, controlling project commands, renaming a process, opening terminals, sending input, or configuring restart behavior. Do not use for read-only status and readiness, project or solo.yml setup, or failure diagnosis.

Create, inspect, update, delete, fill, and export global or project Solo prompt templates with reusable placeholders. Use when saving recurring prompts, standardizing worker briefs or handoffs across agent tools, managing template scope, or backing templates up as Markdown. Do not use for current task state or durable project decisions.

Create and maintain project-scoped Solo todos with status, priority, tags, blockers, comments, locks, completion, and cross-project transfer. Use when turning plans into tasks, finding ready work, claiming todos, recording progress or handoffs, updating dependencies, or completing work. Do not use for narrative context, which belongs in scratchpads, or for controlling processes.

Build a signed, notarized macOS .app and publish a GitHub release. Use when cutting a new version of WisprDuck.

Use this skill whenever the user wants to generate many AI songs, build or operate an AI radio station, turn listener requests into music prompts, use Cloudflare Workers AI with MiniMax Music, or coordinate endless/batch song generation with Queues, R2, Durable Objects, Cron, or Workflows.

Land a PR by monitoring conflicts, resolving them, waiting for checks, and squash-merging when green; use when asked to land, merge, or shepherd a PR to completion.