AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
davidsimoes avatar

davidsimoes

11 listings · 0 installs

Open-source publisher. Listings imported from github.com/davidsimoes — credited to the original author with their license.

↗ github.com/davidsimoes
11 results
Self-run
SKILL

Clarify

Force a self-audit of the current request — list every ambiguity, assumption, unilateral tradeoff, and irreversible-action risk, then ask the user via AskUserQuestion before proceeding. Use when the user says '/clarify', 'any doubts?', 'are you sure?', 'what are you assuming?', 'check with me first', 'wait — confirm', 'before you start', or any time you catch yourself thinking 'I'll assume X' / '…

0
39
Free
Self-run
SKILL

Restart Process

Capture every in-flight tmux Claude session into an atomic on-disk manifest BEFORE a reboot or system update, so /restart-resume can spawn fresh Claude processes post-reboot with full handoff context as the first user message. Use when you say 'restart my computer', 'reboot and update', 'wrap everything for reboot', 'preserve sessions across reboot', 'capture all sessions for restart', 'safely cl…

0
35
Free
Self-run
SKILL

Pre Send

Pre-send approval ritual that creates the flag required by the send-gate.sh PreToolUse hook. Run this BEFORE any external-send tool (Slack, Gmail, Missive, WhatsApp, HubSpot engagement, DocuSign, etc.). Shows the final draft, requires explicit 'send it' confirmation, then drops /tmp/claude-send-approved with 300s TTL. Triggered automatically by Claude when about to call a send tool, or manually b…

0
30
Free
Self-run
SKILL

Respond Html

Render a structured HTML artifact instead of a chat reply — for plans, proposals, comparisons, audits, recommendations, mockups, and process/data visualizations. Single self-contained file saved under an html-answers/ directory, opened in browser, with inline ✅/💬/❌ reactions + comment box on every section, decision-block, and callout. Live stats in sticky toolbar. One button copies all feedback…

0
34
Free
Self-run
SKILL

Chrome Validate

Browser-based QA validation for any URL or PDF — screenshots, computed CSS, layout integrity, link liveness, network inspection, PDF text-diff with diacritic check. Use when validating a deck/prototype/PDF before sending to clients, verifying visual fixes via Chrome MCP, checking that font subsetting didn't eat non-ASCII characters (e.g. Czech/Polish/Vietnamese diacritics), or saying 'validate th…

0
44
Free
Self-run
SKILL

Dataviz

Visualizes any data OR process as a client-ready interactive HTML report. DATA MODE: statistician-grade EDA from JSON/CSV — profiles, extracts signals, picks the right viz library (Observable Plot, Plotly, Vega-Lite, Chart.js). PROCESS MODE: visualizes workflows, pipelines, architectures, state machines — picks Mermaid or rich custom HTML with step cards. Use on explicit /dataviz invocation, or w…

0
35
Free
Self-run
SKILL

Fresh Eyes

Hostile audit of current plan or implementation. Spawns parallel subagents to verify everything against reality. Produces Reality Score.

0
29
Free
Self-run
SKILL

Delegate

Spawn and cooperate with a child Claude session for a specific task. Handles tmux placement (pane/window/session/worktree), task-prompt injection, registry tracking, and returns a clean handle for follow-up. Cooperation verbs: /delegate list|status|tell|fetch|close <name>. Use when you want to hand off a scoped task to a separate session, say 'delegate this to another Claude', 'spawn a coach/agen…

0
33
Free
Self-run
SKILL

Second Opinion

Multi-model review of a specific artifact (file, doc, strategic plan, PRD, decision). Runs Gemini + Claude Sonnet + Claude Opus in parallel — three independent reviewers with different strengths. Use when: 'get a second opinion on X', 'review this doc from multiple angles', 'multi-model audit of this plan', 'second opinion before I ship this', 'what would a hostile reviewer say'. Distinct from /f…

0
27
Free
Self-run
SKILL

Plan Archive

Archive the current planning session or a specific plan to a dated archive directory. Preserves thinking evolution so AI can go back and evaluate how architecture decisions evolved. Use when: 'archive this plan', 'save this planning session', 'log this architecture decision', 'preserve this approach', 'store this plan for retrospective'. Complements ExitPlanMode — use immediately after exiting pl…

0
37
Free
Self-run
SKILL

Restart Resume

Restore every tmux Claude session captured by /restart-process — recreates the original tmux layout (sessions, windows, panes with correct geometry via window_layout strings) and spawns FRESH claude processes (NOT --resume) with each pane's captured handoff as the first user message. Use when you say '/restart-resume', 'resume sessions after reboot', 'restore my tmux setup', 'bring back my claude…

0
41
Free
You've reached the end · 11 loaded