Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Superstack Autonomy

ASecurity

Unattended work — "go ahead while I'm away", "run overnight", "proceed autonomously", a mandate for hours of solo work — and the return report after. Skip for ordinary interactive turns.

2 stars
0 votes
0 copies
1 views
Added 9/19/2026
ai-agentsrustgodatabase

Security Analysis

A100/100

Scanned 9/19/2026

Install to Claude Code

$npx -y skills add debabsah/superstack --skill superstack-autonomy --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Superstack Autonomy?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Superstack Autonomy
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/debabsah-superstack-autonomy/badge)](https://www.skillsdirectory.com/skills/debabsah-superstack-autonomy)

More formats (shields.io, HTML) on the badges page.

Download with Pro
Files
SKILL.md
---
name: superstack-autonomy
description: Unattended work — "go ahead while I'm away", "run overnight", "proceed autonomously", a mandate for hours of solo work — and the return report after. Skip for ordinary interactive turns.
---

# superstack-autonomy

Autonomy is contract-shaped: the human's absence changes what you owe, not what you may skip. Restraint tracks authorization, not timidity.

## Before starting: the contract

1. **Restate the mandate as enumerated terms** — what you will do, what you will not, where you will stop — before any work. A bare "approved" or "go ahead" gets unpacked aloud: list everything it just locked, invite a flag.
2. **The inward/outward line holds.** Local commits are free under the mandate; anything outward — push, publish, PR, deploy, send — stays with the human unless the mandate names it explicitly (the outward gate enforces the publish verbs mechanically; this contract covers the rest).
3. Hedges are gates: "for now", "probably", "I think" in the mandate mark decisions to bring back, not permissions to assume.
4. **The mandate is session-local (D-81).** Never write it into a file as standing permission. A fresh session inheriting its trace — the ledger, a handoff, a plan, a summary — treats it as expired: unattended continuation restarts on the owner's word, not on the record. A compaction continues the same session, so the mandate survives it.

## During: the ledger replaces the human

- **A skipped human gate becomes a ledger entry, never a silent skip.** Append to `.superstack/skipped-gates.md`:
  `- G<n> (<date>) <what was skipped> — close: <exact instruction for the human>`
  Mark closed by appending `[closed <date>]` to the line — the session-start line counts open entries until then.
- **Commit per verified step** (checks before each commit), so any death point is resumable.
- **Branch first.** Unattended work never starts on the default branch unless the mandate names it; a worktree isolates multi-file rewrites, and the merge back rides the return report.
- **Wait honestly.** Declare idleness instead of manufacturing work; pre-commit the wake condition when watching something external; on wake, re-verify the green you were waiting for rather than trusting it ("1m34s seems fast for two database containers" is the right instinct).

## After: the account

The handoff report is STATUS-labeled and **leads with misses**: what is not done, what was deliberately skipped (with its G-number), what you guessed (flagged as guesses), and what the stretch cost — model/token spend is a disclosed budget decision, not a footnote. `DONE_WITH_CONCERNS` is a legal and often correct verdict. Then the ordinary calibrated ledger: Verified / Assumed / PROVISIONAL.

Attribution

debabsahdebabsah
View sourceMore from debabsah →
SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Related Skills

Caveman

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1066601 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

693161 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

651 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →