CtrlK
BlogDocsLog inGet started
Tessl Logo

global-agent-guardrails

One shared denylist of catastrophic shell commands (rm -rf on / or ~, dd/mkfs, sudo rm, fork bombs, curl|sh, git push --force, gh repo delete) enforced as a PreToolUse/pre-exec guard across every AI coding agent on the machine — Cursor, Claude Code, Codex, OpenCode, Pi, Hermes, Grok, Droid, Devin. Use when adding or tuning blocked-command patterns, wiring the guard into a new agent or a new machine, debugging why a command was (or was not) blocked, or when the user mentions command guard, guardrails, dangerous command hook, or PreToolUse safety.

77

Quality

96%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No known issues

SKILL.md
Quality
Evals
Security

Quality

Content

92%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-sequenced body with strong validation checkpoints and no conceptual padding; its only weakness is structure — it is a long monolithic file with no offloading of the per-agent wiring or gotchas into bundle reference files.

Suggestions

Move the per-agent wiring table and/or the Gotchas list into a references/ file (e.g. references/wiring.md, references/gotchas.md) and signal them with one-level-deep links from the body to improve progressive_disclosure.

Consider extracting the agent-specific adapter snippets (OpenCode/Pi/Hermes) into a scripts/ or references/ bundle so the SKILL.md body stays a concise overview.

DimensionReasoningScore

Conciseness

Dense and fluff-free with no explanation of concepts Claude already knows — every section is operational (file map, wiring table, gotchas, verification recipe). Not a 2 because there is no unnecessary explanation to tighten; the volume reflects the genuine 10-agent scope rather than padding.

3 / 3

Actionability

Provides exact file paths, copy-paste JSON hook entries, concrete shell test commands, and a direct script probe. The /ABSOLUTE/HOME/ placeholder is explicitly justified ("Use absolute paths in configs"), so it stays copy-paste ready. Not a 2 because the guidance is fully executable, not pseudocode.

3 / 3

Workflow Clarity

The "Add or tune a pattern" section is a clear numbered sequence with explicit validation checkpoints ("Must pass 100%", the python compile check, test-guard.sh ending "failed: 0") and an E2E verification recipe. Not a 2 because validation and feedback steps are explicit; the destructive-operation cap does not apply since verification is present.

3 / 3

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ are absent) and the 95-line body is a monolith with substantial inline content (the per-agent wiring table and the gotchas list) that could be split into reference files. Not a 3 because there are no signaled one-level-deep bundle references; not a 1 because sections are well-organized, not a nested wall of text.

2 / 3

Total

11

/

12

Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description with an explicit Use-when clause, concrete action list, and natural trigger terms spanning the skill's distinct niche. Voice is third person / imperative with no first- or second-person phrasing, so no voice penalty applies.

DimensionReasoningScore

Specificity

Names multiple concrete actions — "adding or tuning blocked-command patterns, wiring the guard into a new agent or a new machine, debugging why a command was (or was not) blocked" — plus an explicit command denylist (rm -rf, dd/mkfs, curl|sh, git push --force). Matches the anchor listing multiple specific concrete actions; not a 2 because it is comprehensive rather than partial.

3 / 3

Completeness

Explicitly answers what ("One shared denylist of catastrophic shell commands ... enforced as a PreToolUse/pre-exec guard") and when via an explicit "Use when ..." clause with concrete triggers. Matches the anchor requiring both what and when; not a 2 because the when clause is explicit, not merely implied.

3 / 3

Trigger Term Quality

Includes natural phrasings a user would say — "command guard, guardrails, dangerous command hook, or PreToolUse safety" — covering common variations. Not a 2 because it spans several natural synonyms rather than a single keyword.

3 / 3

Distinctiveness Conflict Risk

A clear niche (cross-agent catastrophic-command guardrails) with distinct triggers unlikely to overlap with other skills. Not a 2 because the scope and trigger terms are specific enough to avoid collisions.

3 / 3

Total

12

/

12

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
mishatojk/davidskills
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.