Skip to content
Back to skills

Execution Integrity

ASecurity

Seal exact-byte run packets and preserve reservations, failed attempts and resource costs across repairs or host changes. Use for bounded expensive jobs, staged readiness or evidence handoffs; not for ordinary edits needing no execution contract.

  • 6 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 11, 2026
ai-agentsnode

Security analysis

A100/100

Pro scans all 6 files and shows the line behind each finding

Scanned November 11, 2026

npx -y skills add TianMingKanae/engineering-multi-ai-skills --skill execution-integrity --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Execution Integrity?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Execution Integrity
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tianmingkanae-execution-integrity/badge)](https://www.skillsdirectory.com/skills/tianmingkanae-execution-integrity)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: execution-integrity
description: Seal exact-byte run packets and preserve reservations, failed attempts and resource costs across repairs or host changes. Use for bounded expensive jobs, staged readiness or evidence handoffs; not for ordinary edits needing no execution contract.
---

Identify the authorized phase, accountable implementer, input inventory, shared budget and stopping condition. Preparation, stored-data compatibility, execution permission and accepted outcomes are different facts. A preparation receipt never releases an execution job. Use [packets.md](references/packets.md) to seal and check a declared file inventory, and [budgets.md](references/budgets.md) when durable attempt reservations are needed.

Resolve the runtime and imported dependencies on the target node, recording actual paths, versions and hashes in a local receipt. A matching source file or stored-data check does not establish runtime compatibility. Keep each active run's source/configuration snapshot immutable. Repairs run from a fresh root; preserve the old error, outputs and charged attempts. Hash-pinned configuration and evidence need exact-byte transport, including line endings and encoding.

Reserve bounded work before launching. Keep pending/crashed reservations charged until an evidenced settlement; retain failed-attempt costs. Distinguish measurements from conservative policy debits and estimates. Count distinct process lifetimes once, and do not sum a parent interval with its nested diagnostics. Host-wall totals can exceed elapsed wall time when hosts overlap. Neither included helper launches work, measures a process tree, authorizes spending or implements distributed consensus.

Complete ordinary authorized repairs without adding administrative handoffs. Recheck affected invariants and send one scoped delta packet at the applicable milestone; reuse unchanged accepted evidence. A changed criterion, workload, dependency semantics or resource ceiling requires a prospective decision, not a mechanical-fix label.

Report the narrowest supported outcome: declared bytes intact, preparation checked, command finished, criteria passed or independently verified. Name missing journals, unchecked behavior and historical unmeasured costs. Never turn an intact packet, author test or model exit code into a VERIFIED claim.

Files in this skill

  • SKILL.md2.3 KB
  • agents/openai.yaml240 B
  • references/budgets.md4 KB
  • references/packets.md3.6 KB
  • scripts/budget.py6.5 KB
  • scripts/packet.py4.9 KB

Attribution

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

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.