IgorGanapolsky
UserThumbGate Pre-Action Checks self-improve from ranked lessons and repeated failures, hard-block detected secret leaks, and block matches in strict mode.
Categories
Indexed Skills (93)
evidence-first-answer
Required pre-flight before answering any "why" or "are we" question about revenue, conversion, outbound, traffic, or commercial state. Forces verification of live data and distinguishes planning docs from telemetry. Triggered automatically when the CEO asks "are you sure?" — that phrase means the previous answer was wrong.
revenue-truth
Fetch live ThumbGate revenue, paid orders, and funnel telemetry from production billing endpoints. Use BEFORE answering any question about money, conversion, traffic, or "why we didn't make X." Never quote planning documents as actuals.
autonomous-execution-mandate
Self-directed autonomous execution mandate and workflow skill for ThumbGate agents. Enforces zero-human-intervention task chaining, proactive CI/Trunk monitoring, automated conflict resolution, and evidence-first reporting without requiring manual prompt nudges from the user.
gsd-ralph-context-loop
Orchestrate coding work with GSD (Capture→Clarify→Organize→Execute→Review) and Ralph Loop (Observe→Act→Feedback→Promote→Enforce). Use parallel agents/workflows for multi-faceted tasks. Auto-invoke when user says GSD, Ralph Loop, get shit done, work in parallel agents, context course, or all of it for context engineering. Slash: /gsd-ralph-context-loop.
thumbgate
ThumbGate provides pre-action gates for AI coding agents. It captures thumbs-up/down feedback on agent actions, auto-promotes repeated failures into prevention rules, and blocks known-bad tool calls via PreToolUse hooks. Trigger when the user wants to add safety guardrails to an AI agent workflow, capture structured feedback on agent output, generate prevention rules from failure patterns, gate high-risk actions before execution, or export DPO training pairs from production feedback. Works with any MCP-compatible agent including Cursor, Codex, Gemini CLI, Amp, and OpenCode.
skool-headless-reader
Read, monitor, and analyze Skool communities or posts headlessly for lead discovery, customer pain, acquisition opportunities, or ThumbGate revenue research without blocking the user's browser.
ci-failure-triage
Triage a CI / PR check failure by READING the failure body before forming any hypothesis. Triggered whenever a required check is red, a PR is BLOCKED, a merge won't land, or you're about to call a failure "transient", "flaky", "stale", or "orphaned". Prevents dismissing a real failure (e.g. real CodeQL security alerts) as noise.
scope-discipline
Keep one task closed-looped before starting the next. Triggered when a new problem surfaces mid-task (a bug, a security finding, a missing env var), when you're about to branch into unrelated work, or when you notice the session has sprawled across many half-finished threads. Prevents the "nothing ever lands" failure where each step spawns new work and none finishes.
task-state-ledger
Maintain an external state ledger for any multi-step task (3+ steps, multi-PR, anything spanning more than one turn or touching production). Triggered when you notice yourself re-deriving "where am I" with repeated gh/git/curl dumps, when a task crosses turns, or when the user asks "why are you stuck/lost". Prevents the "lost" failure mode: re-discovering state every turn instead of reading it.
agent-memory
Give your agent persistent cross-session memory — recall what went wrong on similar tasks before starting, then capture whether the work succeeded or failed so repeated mistakes auto-promote into prevention rules. Runs locally via the ThumbGate MCP server (recall, capture_feedback, prevention_rules, feedback_stats) — no API key needed. Use when the user says "recall", "what went wrong last time", "remember this for next time", "capture feedback", "thumbs up/down on that", or at the start of a new task or session. Do NOT use when the user wants to force one specific mistake into an enforced block gate (use the thumbgate-guard skill), to list rules that already exist (thumbgate-rules), or to check ThumbGate wiring (thumbgate-doctor).
bluesky-engagement
Automated Bluesky reply monitoring and draft queueing for ThumbGate's acquisition engagement loop. Polls the AT Protocol notifications endpoint every 15 minutes, writes human-reviewable draft replies into a queue file, and never auto-posts without sign-off. Trigger when the user asks about Bluesky replies, engagement, reply monitoring, or wants to see/approve queued drafts. Also the authoritative reference for rotating the Bluesky app password or diagnosing monitor failures.
thumbgate-blocked
Show ThumbGate's enforcement record — how many risky actions were actually blocked versus warned, which gates fire most, the tokens/damage saved, and the full feedback to check to rejection pipeline. Reads live enforcement counters via the gate_stats and enforcement_matrix MCP tools (CLI fallback `npx thumbgate gate-stats`). Use when the user says "what has ThumbGate blocked", "show gate stats", "is enforcement working", "how many tokens did we save", "show the enforcement matrix", or "what got stopped". Do NOT use to list rule DEFINITIONS that exist (use the thumbgate-rules skill), to create a new rule (use the thumbgate-guard skill), or to check whether ThumbGate is installed and wired (use the thumbgate-doctor skill).
thumbgate-doctor
Health-check whether ThumbGate is actually wired into this agent — PreToolUse/SessionStart hooks installed, MCP server reachable, lesson store present, statusline, and overall agent-readiness — then report exactly what to fix. Runs the existing `npx thumbgate doctor` audit and the check_operational_integrity MCP tool. Use when the user says "is ThumbGate wired up", "thumbgate doctor", "check my guardrails are installed", "why aren't my gates firing", "is the MCP server connected", or "agent readiness". Do NOT use to view rules (use the thumbgate-rules skill), to view what was blocked (use the thumbgate-blocked skill), or to capture a new rule (use the thumbgate-guard skill) — this skill only diagnoses setup and wiring.
thumbgate-guard
Turn the agent's most recent mistake into an enforced ThumbGate prevention rule (a PreToolUse block gate) so the same bad tool call is intercepted before it runs again, in this and every future session across Claude Code, Cursor, Codex, Gemini, Amp, and Cline. Captures the failure with the capture_feedback MCP tool, then force-promotes it via `npx thumbgate force-gate` so it is enforced, not just logged. Use when the user says "guard against this", "block this from happening again", "never do that again", "make that a rule", "stop the agent from repeating that", or right after a bad action or thumbs-down that should become a hard rule. Do NOT use to merely log a thumbs-up/down without enforcement (use the thumbgate-feedback skill), to recall prior context before starting work (use the Agent Memory skill), or to list rules that already exist (use the thumbgate-rules skill).
thumbgate-protect
Inspect this repo's branch and release governance (protected branches, release rules, protected-file globs) and, only when the user explicitly approves, grant a scoped, time-limited exception so a protected-file edit or publish can proceed under audit. Reads posture via the get_branch_governance MCP tool and records a narrow, expiring approval via the approve_protected_action MCP tool. Use when the user says "is main protected", "show branch governance", "what am I blocked from editing", "approve this protected change", or "let me edit a protected file just this once". Do NOT use to disable protection wholesale, to grant broad or standing exceptions, or to diagnose hook wiring (use the thumbgate-doctor skill) — this skill is for narrow, temporary, audited approvals only.
thumbgate-rules
List the active ThumbGate prevention rules, reliability rules, and the promoted lessons behind them, so the user can see which guardrails are currently protecting this project and WHY each one exists. Reads the live rule and lesson stores via the prevention_rules, get_reliability_rules, and search_lessons MCP tools (CLI fallback `npx thumbgate rules`). Use when the user says "what is ThumbGate protecting me from", "show my rules", "show my gates", "what has the agent learned", "list active guardrails", or "what's blocked here". Do NOT use to CREATE a new rule (use the thumbgate-guard skill), to see runtime enforcement counts of what actually fired (use the thumbgate-blocked skill), or to diagnose whether ThumbGate is wired up at all (use the thumbgate-doctor skill).
context-engineering-checklist
HARD checklist before new skills, MCP tools, hooks/gates, plugins, or multi-agent workflows. Maps Hugging Face Context Course units 1–5 onto ThumbGate. Auto-invoke when creating SKILL.md, PreToolUse hooks, gate patterns, MCP tools, Grok workflows, slash commands, or when user says context engineering, HF context course, GSD, Ralph Loop, or agent context. Slash: /context-engineering-checklist.
agent-harness-router
Agent harness router consolidating TrueFoundry TrueForge (harness as control point, $10/mo budget enforcement), Martech (scattered AI discovery capture), GLM-5.3 (private benchmark validation, model portability), and Futurism (ChatGPT Computer History fail-closed alternative). Routes model calls by cost/latency/quality/privacy, tracks spend against hard cap, captures reusable prompt patterns, validates models on private benchmarks. Never captures keystrokes, clicks, or local Mac activity.
live-dashboard-verify
Verify live ThumbGate.app service health, trace header propagation, and production deployment status using curl + BrowserOS.
prolo-ring-moodstrip
Autonomous MoodStrip setup for Prolo Ring Bluetooth media controls. Configures Play/Pause, Next, Previous, Volume controls via Bluetooth HID. Handles pairing, flashing, and connection monitoring.
revenue-gap-tracker
Track revenue monitoring gaps on ThumbGate.app - identify blocked data sources (gitignored business_os/, admin auth, Stripe env vars), audit available signals, and determine what is needed to move revenue monitoring from 3 to 5.
xurl
X/Twitter via xurl CLI: post, search, DM, media, v2 API.
solve-architecture-autonomy
Course-correct from over-planning to execution — when the agent has produced a plan, research, or options instead of doing work the user already approved, drop the analysis and execute the smallest end-to-end action now (branch, run, deploy, verify). Use when the user says "stop planning, just do it", "why are you researching, execute", "just ship it", "enough analysis", or pushes back that you keep proposing instead of acting on an already-decided task. Do NOT use when the task is genuinely undecided and needs a plan first, when the user explicitly asked for research/options/a design, or for routine work where no over-planning has occurred — acting on an unscoped or destructive task is its own failure.
ci-pr-monitor
Repeatable workflow for monitoring CI check status, tracking PR merge-readiness, and verifying auto-merge success using the gh CLI.
adversarial-review
Run a skeptical Codex review from Claude Code that hunts for hidden regressions, unsafe assumptions, and release risk. Use before merges, deploys, migrations, or other risky changes.
review
Run a Codex review from Claude Code against the current uncommitted work, a base branch diff, or a specific commit. Use when the user explicitly asks for a Codex review or second reviewer.
second-pass
Hand the current task or repo state to Codex for an independent second pass from inside Claude Code. Use when the user explicitly wants another agent to take a shot after Claude's first pass.
capture-feedback
Capture structured thumbs up/down feedback with context, history-aware lesson distillation, tags, and optional rubric scores after completing a task.
prevention-rules
Generate and review prevention rules auto-promoted from repeated failure patterns.
programmatic-agent-runs
Govern Cursor SDK local, cloud, self-hosted, and subagent coding runs before they create branches or PRs.
recall-context
Recall relevant past failures, prevention rules, and context packs before starting a coding task.
search-lessons
Search promoted lessons for corrective actions, lifecycle state, linked rules, and linked gates.
graphify
Query the local Graphify-Labs AST knowledge graph for ThumbGate architecture questions (what connects X to Y, explain a symbol, PreToolUse/gates/adapters). Slash: /graphify.
cyberstrike-compare-not-clone
CyberStrikeAI (Ed1s0nZ) turns NL intent into governed cybersecurity actions with HITL + RAG. Steal the FORMAT onto ThumbGate rails; never clone CyberStrike, Eino, WebShell, or C2. Slash: /cyberstrike-compare-not-clone.
jit-harness-compare-not-clone
JIT-Agent (arXiv:2608.25593) is a harness-intelligence model, not a ThumbGate clone. Steal the four-module FORMAT (memory/planning/action/capability) onto existing rails; never train or download JIT-Agent. Slash: /jit-harness-compare-not-clone.
nvidia-specdecode-al-doctor
Fail-closed speculative-decoding AL/D doctor for ThumbGate. Checks speedup ≤ AL/(1+ρD), attention D=128/G-1, tile alignment, and maps to checkpoint-speculative-decoding-acceptance. Steal from NVIDIA co-design blog — never clone TensorRT/EAGLE/Model-Optimizer. Slash: /nvidia-specdecode-al-doctor.
show-me
Answer architecture / control-flow questions with a compact visual format (component tree, call stack, mermaid, annotated diff) instead of prose walls. House steal of ExplainX /show-me FORMAT — not affiliated. Slash: /show-me.
zvec-grep-compare-not-clone
zg (zvec-grep) is a local-first search layer (ripgrep + BM25 + vector), not a ThumbGate clone. Steal the four-route FORMAT onto existing rails; never install @zvec/zvec-grep. Slash: /zvec-grep-compare-not-clone.
thumbgate-feedback
Capture thumbs feedback and apply prevention rules before coding
result
Print the latest saved Codex bridge result from Claude Code without rerunning Codex. Use when the user asks for the last Codex output or wants to inspect the raw bridge result.
setup
Validate that the Codex bridge plugin is ready inside Claude Code. Use when the user asks to install, configure, or sanity-check Codex review from Claude.
package-manager-honesty-doctor
Audit lockfile/packageManager/CI install parity and fail-closed package-manager switches. Steal from InfoQ pnpm 12 (Rust rewrite) — do NOT migrate ThumbGate off npm. Prefer --ignore-scripts honesty. Slash: /package-manager-honesty-doctor.
dogfood
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.
advertise-thumbgate-github-marketplace
Cite the live GitHub Marketplace Action listing when advertising ThumbGate (free listing only). Trigger: advertise ThumbGate on Marketplace, listing badge, uses: snippet for README, Marketplace install URL, Action discovery. Slash: /advertise-thumbgate-github-marketplace. Never claim LIVE without /verify-github-marketplace-listing. Never paid plans or buyer outreach without /eci-thumbgate-ip-wall clearance.
anti-babysitting
Igor's standing directive (2026-08-10): never end a turn asking him for approval, a click, or a decision the agent could make. Use at the START of any autonomous run, and any time you are about to write "let me know", "one click away", "should I", or "needs your approval" — those phrases are the trigger. Defines the pre-authorized action classes, the exhaust-then-park protocol for platform walls, and the short list of actions that still require explicit consent.
elevenlabs-voice-agents
ElevenLabs Hosted MCP Voice Agents, Conversational AI Engine, Multi-LLM Cost Estimator & ThumbGate Pre-Action Interdiction Suite for all coding agents on Igor's Mac.
github-marketplace-action-publish
Publish or update a GitHub Actions listing on GitHub Marketplace (free only for ThumbGate until counsel clearance). Trigger: GitHub Marketplace, publish this Action, Draft a release, marketplace listing, advertise the action, uses: IgorGanapolsky/ThumbGate, action.yml branding. Slash: /github-marketplace-action-publish. Never claim LIVE without /verify-github-marketplace-listing. Never add paid Marketplace plans without /eci-thumbgate-ip-wall clearance.
gurobi-harness-gate
Session/process gate: Gurobi certifies $10 token-budget LP and exclusive adb/git_lock dispatch at session-start and agent-loop health. Proof receipts, not LLM guesses. Trigger: gurobi harness, session start proof, token budget gate, exclusive dispatch, certified allocation.
gurobi-optimizer-integrator
Real Gurobi (gurobipy) optimization for fleet scheduling, AHLS outreach batch selection, LP/MIP, IIS infeasibility proof, and $10/mo token-budget LP. Free pip size-limited license first (AE Fabrizio Ellis 2026-08-12). Trigger: gurobi, optimize schedule, mathematical optimization, dispatch agents, outreach capacity knapsack, IIS, token budget, certified proof.
intent-contract
HARD: durable product intent linked to code and tests (Tieline-inspired). Auto-invoke before product coding, before claiming done/shipped, when ACs drift, blast radius, "what implements this", or user stories vs code. Slash: /intent-contract. Not affiliated with knoxgraeme/tieline.
linear-agent-skills
Complete skill suite for Linear AI Agent workspace automation, multi-agent tagging, cycle-time telemetry, automated project health updates, and ThumbGate RAG retrospective promotion.
outcome-owned-agent-pattern
Grok Bot–style outcome ownership for Hermes/Grok multi-agent fleet: named roles, draft-first boundaries, skill→test→routine ladder, narrow event triggers, no-stale-data. Source: https://docs.x.ai/grok-bot/overview (2026-08 steal). Trigger: steal grok bot ideas, bot charter, skill then routine, draft first automation, outcome-owned agent, teach a task. Slash: /outcome-owned-agent-pattern.
verify-github-marketplace-listing
Evidence-only check whether a GitHub Actions Marketplace listing is actually LIVE. Use when: is it on Marketplace, advertise the action, listing URL, marketplace 404, did we publish the Action, ThumbGate Agent Governance, uses: IgorGanapolsky/ThumbGate. Slash: /verify-github-marketplace-listing. Never claim LIVE from a release title or repo banner alone.
blocked-run-still-ships
What a content or outreach run must still deliver when publishing turns out to be impossible, and how to keep every claim honest. Use when a scheduled ThumbGate content run, social fan-out, or longform syndication hits a credential/OAuth/browser blocker; when about to log a channel as Blocked; when writing a run's PR body, ledger, or operator notification; when generating a meme and Higgsfield has no credits; or when a fact you asserted earlier fails to reconfirm on re-fetch. Reach for this whenever you are tempted to end a run with only blocker notes, to soften a claim you can no longer verify, or to invent some publishable surface just so the run has a URL to show. Pairs with scheduled-publish-preflight, which diagnoses whether publishing is possible in the first place.
scheduled-publish-preflight
Preflight for any scheduled, unattended, or cron-fired publishing run (ThumbGate content engine, social fan-out, longform syndication, outreach sends) BEFORE drafting or posting anything. Establishes the real date, detects whether this session is on Igor's Mac or an ephemeral cloud container, probes publish capability for real, and checks today's other runs for a duplicate beat including UNMERGED branches and open PRs. Use this whenever a run is about to write a content log row, pick a persona/beat, choose a branch or campaign id, or record a channel as Blocked — and especially when you are tempted to reuse a date or a blocker reason from a previous run instead of re-deriving it. Pairs with the blocked-run-still-ships skill, which governs what a run must produce once this preflight says publishing is impossible.
apollo-io-sales
Use Apollo.io for sales contact discovery and email enrichment on Igor's Mac. CLI: /opt/homebrew/bin/apollo (apollo-io-cli), already authenticated. Trigger when: make money, outreach, prospects, enrich email, find founder email, no public email, pipeline send, Partner Pilot, cold call, leads, Apollo, or when tempted to give up because leads.csv says "no public email." Slash: /apollo-io-sales. NEVER skip Apollo when money depends on a real inbox.
drive-logged-in-chrome
Drive Igor's already-authenticated Google Chrome via AppleScript for login-walled dashboards (Stripe Payment Links, Apollo web, Play Console, LinkedIn, ASC). NEVER use Playwright MCP as the source of truth for "am I logged in?" — Playwright is a separate browser profile without Chrome cookies. Trigger when: Stripe, payment links, buy.stripe.com, dashboard.stripe.com, "logged into Chrome", login wall, Playwright got login page but user says they're signed in, Apollo web UI, or any task needs the real Chrome session. Slash: /drive-logged-in-chrome. Cross-ref: use-existing-browser-sessions.
execute-revenue-cash-path
Execute the real cash path for mac-yolo-safeguards / ThumbGate: diagnose funnel, Apollo enrich, Gmail send, live Stripe links from Chrome, pipeline-update, never ship theater. Trigger when: make money, make money today, do everything, cash, revenue, sell, close, outreach, pipeline stuck, $0 revenue, Partner Pilot, diagnostic $499, hardening $1500, or when tempted to build another sales tool instead of sending. Slash: /execute-revenue-cash-path. Do NOT invoke for pure product engineering without a revenue ask.
hermes-context-search
DEPRECATED as of 2026-07-24 in favor of `grepai` (see docs/LOCAL-SEMANTIC-CODE-INDEX.md and the `grepai` MCP server in .mcp.json) — a more mature, actively-maintained (1.8k★) equivalent with native MCP integration and Tree-sitter-aware chunking, independently wired up the same day by another agent. Prefer the `grepai` MCP tool for semantic code search going forward. This tool remains as a zero-dependency CLI fallback (no separate binary install, no MCP config) if `grepai` is ever unavailable.
agent-device
Automates Apple-platform apps (iOS, tvOS, macOS) and Android devices. Use when navigating apps, taking snapshots/screenshots, tapping, typing, scrolling, extracting UI info, collecting logs/network/perf evidence, or planning agent-device CLI commands.
ingest-chat-credentials
When Igor pastes ANY credential (password, API key, token, client secret) in chat, immediately ingest it into macOS Keychain, confirm service+account only, never refuse with "don't send passwords / rotate yourself" homework. Auto-invoke on credential paste. Covers App Store Connect Apple ID, Play, Stripe, fleet ENV keys, EAS tokens.
troubleshoot-hermes-mobile-connectivity
Diagnoses Hermes Mobile "can't reach your computer" errors, pairing failures, and misleading connectivity banners. Use when the phone shows unreachable Mac/Hermes, pairing fails (localIp missing), LAN URLs (10.x/192.168.x) fail on cellular/5G, or sends timeout after 60–120s. Repo mac-yolo-safeguards / hermes-mobile.
verify-hermes-mobile-ship
Verifies Hermes Mobile ship claims with evidence — unit tests, continuous E2E, simulator full-suite, device Maestro. Use before saying fixed/shipped/works on device, when user asks "are you sure?", or after UX changes under hermes-mobile/. Prevents false simulator-E2E claims.
thumbgate-app-first-promo
HARD default product narrative for ThumbGate social/promo: ThumbGate.app is the hero product and primary CTA; Hermes Mobile and thumbgate.ai are downstream defaults. Auto-invoke when: promote ThumbGate, post everywhere, fill the gaps, social promo, content engine, Creator Platform Promo, fan-out, LinkedIn/Medium/dev.to/X/Bluesky/ Threads/Reddit/HN distribute, "did we post", Continuity promo, or any plan that still treats store/mobile as co-equal heroes. Slash: /thumbgate-app-first-promo.
hermes-mobile-connect
Facilitates Hermes Mobile ↔ Mac gateway pairing, onboarding, and ThumbGate install when the user does not have ThumbGate.app. Use when a user needs help connecting their phone to their Mac, when the app reports "can't reach your computer", or when ThumbGate is missing and needs a Herdr-style one-line install + web signup path. Repo: mac-yolo-safeguards / hermes-mobile.
coding-context-pack
HARD issue-first coding context (HF Context Course pattern): before product coding load the smallest pack that proves what "correct" is (open GH Issue AC + Linear + e2e gate + skills). Auto-invoke at coding session start, when user says "what should we work on", "focus issue", "context pack", "automate coding context", after ship theater, before claiming done/shipped. Slash: /coding-context-pack.
anti-babysitting-mandate
HARD: never park waiting for "continue" or "want me to?" on pre-authorized repo/machine work. Residual picker + wrap-up lint. Auto-invoke on continue, want me to, should I, be more autonomous, anti-babysitting. Slash: /anti-babysitting-mandate. Policy home for Grok: ~/.grok/skills/anti-babysitting-mandate.
browseros-agent-harness
BrowserOS & BrowserOS Neo Agent Automation Protocol. Automates real local browser sessions, persistent account logins (Skool, GitHub, Stripe, Linear), local file access, and 53 browser automation MCP tools across all coding agents on Igor's Mac.
ci-first-fail
Buildkite-style first-fail CI annotations. Name the failing STEP before reading logs. Missing required contexts → GitHub Status, not a code blame. Do not migrate this public repo off GitHub Actions. Trigger: red CI, BLOCKED PR wall, "is it the runner", first failing step, Buildkite. Slash: /ci-first-fail.
content-lane
Use when drafting public comments or posts. One lane, brief schema, staged artifacts, proof gate. Not a single chat prompt.
garryslist-opportunity-tracker
Garry's List (garryslist.org) continuous intel + membership engage pipeline. Site is a California builder/civic publication (501c4), not a gig marketplace. Scrape public posts, score AI/tech high-value, stage comments/articles. Login: Google SSO iganapolsky@gmail.com; LinkedIn verify for membership. Trigger: garryslist, Garry's List, garryslist.org, scrape garry, engage garry.
git-at-scale
Client-side Git scale hygiene stolen from Cursor Continuity ("Git at any scale"): multi-pack-index, commit-graph, geometric repack, safe agent worktree GC, tip-consistency checks. NOT Origin/WAL hosting. Trigger: packs high, worktree sprawl, slow git, Cursor Continuity blog, git at any scale. Slash: /git-at-scale.
git-hooks-no-husky
Husky process steal without the npm package. Keep core.hooksPath=.githooks (relative) so each worktree runs its own hook. Do not install husky. Trigger: husky, typicode.github.io/husky, hooksPath, worktree hooks, would husky help. Slash: /git-hooks-no-husky.
gurobi-decision-audit
Beyond LLMs process steal: certified Gurobi allocation is not an execute license. Append-only audit + repeatability fingerprint. Human oversight for ADB/spend/send. Trigger: gurobi audit, execute license, decision intelligence, beyond LLMs, governance oversight.
hidden-entry-not-brighttalk
BrightTALK weekly-rec process steal: interest-ranked digest of hidden agent-tool-call entry points (PreToolUse unwired, dynamic tools, missing identity). Do not clone BrightTALK, SailPoint, Strike48, or ISO 42001. Trigger: brighttalk, hidden entry points, going agentic SOC, ISO 42001, interest digest, misconfigurations attackers rely on. Slash: /hidden-entry-not-brighttalk.
hosted-computer-stack
Honesty doctor for Perplexity Computer / OpenClaw / E2B / Cua blueprints and for ChatGPT Computer History / Windows Recall / Mac input-logger asks. We are not those products. Hosted Hermes is $10/mo chat on a fenced VPS. Hands is POLICY_CUE_NOT_DRIVER. History helper is FAIL_CLOSED. Trigger: OpenClaw, Eigent, E2B, trycua, Cua, Agent S, Computer History, Windows Recall, keystroke capture, Manager/Factory/Hands, clone Computer.
hybrid-route-policy
Lemonade-router-inspired hybrid routing for Hermes economic routes: keyword rules, sensitive→local, coding→local candidate, paid cloud only with paid-ok. Trigger: hybrid router, route policy, local vs cloud routing, lemonade router pattern, economic route rules. Slash: /hybrid-route-policy.
local-ai-use-adapt
Adapt AMD Skills' local-ai-use pattern to route image generation, text-to-speech, and speech-to-text through local AI servers to reduce token cost and improve privacy. Maps to ThumbGate.memory tier. Use when asked about local AI, local image, local TTS, local STT, or token reduction.
local-media-lane
AMD local-ai-use pattern adapted for Apple Silicon: route image/TTS/STT to local Lemonade or Ollama when healthy; never claim local media without probe. Chat stays on Hermes economic router. Trigger: local image, local tts, local stt, lemonade, cut media token cost, multimodal local. Slash: /local-media-lane.
ona-last-mile
Place Mac + Hermes Mobile work local-first using Ona/OpenAI Cloud Agent ideas (persist across devices, customer-controlled credentials, reviewable receipts). Default local under $10/mo. We are not Ona Cloud. Trigger: Ona, Gitpod, Cloud Agents, last-mile, persist across phone and Mac.
openai-ultrafast-fleet
Persist OpenAI Ultrafast (GPT-5.6 Sol + service_tier=ultrafast, Cerebras preview) across shells, launchd, and agent harnesses as OPT-IN only. Hard $10/mo cap. Never default jcode-yolo or hermes-yolo to Sol. Complementary to ultrafast-yolo (exact server-tier proof). Trigger: Ultrafast, GPT-5.6 Sol, Cerebras 750 TPS, service_tier=ultrafast, $10 OpenAI budget.
openai-ultrafast-policy
Route latency-critical, high-value work to OpenAI GPT-5.6 Sol Ultrafast only when preview access, contracted pricing, exact server-tier proof, and the shared $10 monthly API ceiling are all satisfied. Use when a user mentions OpenAI Ultrafast, Cerebras-backed Sol, real-time incident response, live support or voice, urgent fraud or transaction review, checkout recovery, or interactive research where latency materially changes the outcome.
output-quality-loop
Use when drafting or reviewing any LLM artifact. Contract, gold examples, rubric, capped repair. Never one-shot slop.
rubrik-mythos-triage
Rubrik Mythos high-ROI steal: AI discovery outruns human review. Multi-pass filter with file-hash checkpoints, trust-boundary prune, tightly-scoped auto-remediation, and Track B packets for secrets/eval/chains. $10/mo fail-closed (local scan, no paid API). Trigger: Rubrik, Mythos, Glasswing, vulnerability triage, what not to automate.
rule-sprawl-not-opensearch
New Stack alerting/retrieval webinar process steal: prevention rules vs eval budget, retrieval latency budget, review volume is not the control. Do not build OpenSearch, PPL, or Unified Alert Manager. Trigger: newstack, alert rules multiplying, rule sprawl, false-positive fatigue, retrieval under concurrent agents, human review vs verified pipelines. Slash: /rule-sprawl-not-opensearch.
screenpipe-activity-local
Screenpipe v2.6.56 process steal (Activity + multi-harness + local-first egress) without installing Screenpipe capture. Uses mac-computer-history semantic events, Activity bands (task/meeting/break), recall, opt-in --egress-ok packs for Cursor / Claude Code / Codex / Grok. Trigger: screenpipe, screenpi.pe, Activity timeline, computer history local, harness context pack, what context leaves. Slash: /screenpipe-activity-local. We are not Screenpipe.
skill-catalog-governance
AMD-skills-style catalog governance for this fleet: skill-card.md (Description/Owner/License), federated sources.yml pins, fail-closed validation. Trigger: skill card, skill catalog, sources.yml, skill governance, amd/skills pattern. Slash: /skill-catalog-governance.
startup-application-helper
Generates startup application content with Truth Guardrails. Maps product value prop to funding opportunity requirements. Use for startup, funding, or grant applications.
tieline-product-contract
Tieline process steal: reviewed source of truth for product behavior and product-level blast radius. Use when asking how Tieline improves our system, before editing offer/landing files, or when a PR may silently change buyer-visible ACs. Slash: /tieline. Not affiliated with knoxgraeme/tieline.
zai-glm53-fleet
Wire Z.AI GLM-5.3 across Hermes, hermes-yolo, LiteLLM :4010, Claude Code / Cline / OpenCode. Default is the GLM Coding Plan ($0 tokens; 5.2 auto-routes to 5.3). Metered /api/paas/v4 and OpenRouter z-ai/glm-5.3 are fail-closed at $10/mo. Trigger: GLM-5.3, z.ai, coding plan, glm-coding, $10 token budget, thinking mode.
prefer-headless-background-ops
Default automation posture for Igor's Mac: CLI and API first, background headless workers, never hijack interactive Google Chrome or macOS focus unless Igor explicitly asked in that message. Use when tempted to open Chrome, osascript, Computer Use, headed Playwright, or install CDP LaunchAgents.
listen-to-chief
Use before any ThumbGate or mac-yolo-safeguards edit. Chief is the driver.
bsky-post
Post to Bluesky via the atproto API (owned posts, replies, threads). Works from cloud sessions — bsky.social is reachable. Use whenever the CEO asks to post/engage on Bluesky.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.