agentleFS
Sign inStart free
googliciusgooglicius/obsidian-steward/skills/guardrails/SKILL.md

guardrails

Skill85 starsChanged 6 months agoMITView on GitHub

Define safety rules to restrict AI from reading, listing, creating, editing, or deleting specific folders or files. Use when the user wants to set up guardrails, access rules, or protect sensitive areas.

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 15 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create new rule files (Use create tool) when adding rules.

    line 18
  2. Distinctive:

    Keep the rule body brief plus the verification query.

    line 22
  3. Distinctive:

    Add it in the rule body, e.g.

    line 30
  4. Very specific:

    With instruction (explicit only): Use this only when the user explicitly requires strict behavior.

    line 87

“Common” is measured: kept by at least 6 in 10 of the 15 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Guardrails Skill

This skill enables you to define safety rules that restrict which folders and files the AI can access. Rules are stored as separate markdown files in Steward/Rules/ — one file per rule.

Rule File Structure

Each rule lives in its own .md file under Steward/Rules/ (flat structure). Each file has:

  1. YAML frontmatter — the rule definition (name, targets, actions, optional instruction, optional enabled)
  2. Documentation — markdown body explaining the rule
  3. Verification query — a ready-to-run command that triggers the rule so users can confirm it works (see below)

Create new rule files (Use create tool) when adding rules. Edit existing rule files (Use edit tool) when updating.

Documentation

Keep the rule body brief plus the verification query.

Verification Query

Include a verification query in each rule's documentation. This is a Steward command that would trigger the rule (and get blocked), so users can quickly confirm the rule works. Format:

  • Format: Commands start with / (slash + space); put the command on its own line
  • Content: A natural-language request that would perform one of the restricted actions on a target (e.g., list, read, create, edit) Add it in the rule body, e.g. under a Verify or Test section.

Frontmatter Format

FieldRequiredDescription
nameYesA short identifier for the rule (e.g., "No secrets access")
targetsYesArray of path patterns (folders or files). Use folder/ for folders, *.ext for file types
actionsYesArray of actions to restrict: read, list, create, edit, delete, move, rename, copy
instructionNoOptional override only when the user explicitly requires strict guidance. Default is no instruction; violations are serialized for the AI to adapt
enabledNoSet to false to disable a rule. Default is true

Example Rule File

Create Steward/Rules/No secrets.md:

---
name: No secrets access
targets: ['Secrets/', '*.key', 'Credentials/']
actions: [read, list, create, edit, delete]
instruction: 'Never read, list, create, or modify files in Secrets/ or *.key'
enabled: true
---

Blocks access to sensitive files (API keys, credentials, Secrets folder).

**Verify:**

/ List files in the Secrets folder

Create Steward/Rules/No private edits.md:

---
name: No private edits
targets: ['Private/']
actions: [edit, delete]
enabled: true
---

Blocks editing and deleting in the Private folder. Violations are serialized for the AI to adapt.

**Verify:**

/ Edit a note in the Private folder

Target Syntax

  • Folder paths: Secrets/ or Private/ — matches the folder and all contents. Use trailing / for clarity.
  • File patterns: *.key — matches files by extension. Use * prefix for glob-like matching.

Instruction vs No Instruction (Default: No Instruction)

  • Without instruction (default): On violation, the error is serialized as a tool result. The AI sees the block message and can try a different approach.
  • With instruction (explicit only): Use this only when the user explicitly requires strict behavior. The instruction is added to tool guidelines, and violations stop processing immediately.

Download now (ZIP)

  • Other versions15 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 15

    1. ai-analyst-lab/ai-analyst★ 304same name
    2. irahardianto/awesome-agv★ 157same name
    3. bolasblack/claude-skills★ 4same name
    4. hajekim/agentic-design-patterns-skills★ 4same name
    5. dfrysinger/skills★ 3same name
    6. adewale/guardrails-skill★ 2same name
    7. Citadel-Cloud-Management/citadel-saas-factory★ 2same name
    8. meaningfy-ws/skillery★ 2same name
    9. antrafa/skills★ 1same name
    10. BrendenKennedy/claude-for-ai-platforms★ 1same name

    5 more not listed

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • Used alongsideProjects with it also use evaluation, prototype and 2 more.

    Other files people use alongside this one.

    Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
    • NVIDIASkill

      evaluation

      Evaluates accuracy of quantized or unquantized LLMs using NeMo Evaluator Launcher (NEL). Triggers on "evaluate model", "benchmark accuracy", "run MMLU", "evaluate quantized model", "run nel". Handles deployment, config generation, and evaluation execution. Always load when validating completed runs or reporting evaluation scores, including standalone launching-evals analysis. Not for quantizing models (use ptq), standalone serving (use deployment), or comparing completed baseline-vs-quantized results (use compare-results).

      Also in 3 projects that use this

    • GoogleSkill

      prototype

      Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.

      Also in 3 projects that use this

    • GoogleSkill

      grill-me

      A relentless interview to sharpen a plan or design.

      Also in 4 projects that use this

    • MicrosoftSkill

      handoff

      Compact the current conversation into a handoff document for another agent to pick up.

      Also in 3 projects that use this

  • In this repository5 other files it gives its agents.

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free