Index of every page on this site: llms.txt
All skills
jeremylongshore avatar

/agent-safety-preflight

@820dfc8
Run it yourself

Generate a local repository risk receipt before Claude Code or other AI-agent edits. Use when the user asks to prepare a repository for agent changes, check for risky hooks or credential-writing automation, or run the /agent-preflight command. Trigger with "run agent preflight", "check this repo before agent edits", or "/agent-preflight".

Use this Skill: https://skilld.dev/gh/jeremylongshore/tons-of-skills-marketplace/agent-safety-preflight

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencesdecision-rules.md

≈293 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Agent Safety Preflight decision rules

Use these rules when the bundled scanner is unavailable or when a human asks how the receipt was classified.

Green

Return Green only when git state is clean and no high-risk automation markers are present in the inspected local files.

Yellow

Return Yellow when any of these are true:

  • Git state is dirty or unavailable.
  • Agent hooks, broad tool permissions, MCP server configuration, or other agent authority surfaces are present without destructive or credential-writing evidence.
  • The scanner cannot inspect enough of the workspace to make a clean Green call.

Red

Return Red and stop before editing when any of these are true:

  • Destructive recursive delete patterns appear in scripts, hooks, commands, or agent-controlled paths.
  • Credential-writing automation appears in an actual environment or secret-assignment context.
  • Shell or network install chains run from agent-controlled paths without a clear local-only safety boundary.

Evidence discipline

Report file paths and short markers only. Do not paste secrets, tokens, private source, environment values, or payment credentials into the receipt.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts12d3 checks · Risk SAFE
  • Gen Agent Trust Hub12d

    The skill is a security-focused utility designed to scan local repositories for risks before AI-agent edits. It performs local analysis using Git and a bundled Python script to identify destructive patterns or credential exposures. While it processes untrusted local data, it includes explicit instructions to avoid exfiltrating sensitive content.

  • Socket12d

    No alerts

  • Snyk12d

    Risk: LOW · No issues

Signed by skilld at 820dfc8. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 months ago
What it can do
Reads files Runs commands
argument-hint
[repo-path]
version
1.2.0
author
Signal Loom Works <194151508+el-zachariah@users.noreply.github.com>
All 3 allowed tools
ReadBash(git:*)Bash(python3:*)
Other metadata
compatibility
Designed for Claude Code
tags
[
  "security",
  "agent-safety",
  "claude-code",
  "preflight"
]

README badge

README badge for jeremylongshore/tons-of-skills-marketplace/agent-safety-preflight