agent-skillguard

agent-skillguard is a skill for Codex from PracticalSwan/agent-skills. It costs 55 tokens per session (1,045 once invoked), scanned A, original, MIT.

A read-only checker for agent skills and plugins before they are installed. It scans local files for risky instructions, excessive permissions, suspicious downloads, prompt-injection patterns, and possible secret exposure.

In plain words
What is it for?
Use it to assess whether an agent skill or plugin should be trusted, installed, or granted permissions. It also records findings, accepted false positives, and remediation steps.
Why use it?
It helps you review an unfamiliar add-on without running or uploading it. Its report points to the exact file and line, while making clear that a scan cannot prove an add-on is safe.

Skill for Codex

Written for Codex: reads ~/.codex or $CODEX_HOME. Also seen: mentions Claude Code; mentions Codex.

Good fit Use it to assess whether an agent skill or plugin should be trusted, installed, or granted permissions. It also records findings, accepted false positives, and remediation steps.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/practicalswan/agent-skills/agent-skillguard
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add PracticalSwan/agent-skills --skill agent-skillguard
Clone the repo
git clone --depth 1 https://github.com/PracticalSwan/agent-skills

Made for: Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for agent-skillguard

README.md
[![agentmods](https://agentmods.dev/badge/skills/practicalswan/agent-skills/agent-skillguard/github.svg)](https://agentmods.dev/skills/practicalswan/agent-skills/agent-skillguard)
Your own site
<a href="https://agentmods.dev/skills/practicalswan/agent-skills/agent-skillguard"><img src="https://agentmods.dev/badge/skills/practicalswan/agent-skills/agent-skillguard/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for agent-skillguard

Your own site · 80×15
<a href="https://agentmods.dev/skills/practicalswan/agent-skills/agent-skillguard"><img src="https://agentmods.dev/badge/skills/practicalswan/agent-skills/agent-skillguard.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 55 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,045 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00055 $0.01045
Opus 5 $0.00028 $0.00522
Sonnet 5 $0.00011 $0.00209
Haiku 4.5 $0.00006 $0.00104

Measured 4d ago against content hash 44a58ff54666, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

agent-skillguard scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

The scan reads SKILL.md. This mod also ships 6 executable files (fixtures/negative/narrow-delete.ps1, fixtures/positive/sg002.sh, fixtures/positive/sg005.ps1, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agent-skillguard/SKILL.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill Risk Check

Use this skill when the user asks whether an agent skill or plugin should be trusted, installed, reviewed, or admitted.

Non-negotiable boundary

Scanning is read-only. Never execute, source, import, install, or enable the target artifact during review. A clean report is not proof that an artifact is safe, and a finding is not proof of malicious intent.

Workflow

  1. Identify the exact local target and its provenance.
  2. Run python <catalog-root>/agent-skillguard/scripts/skillguard.py scan <path> --format markdown before any installation step. Use the absolute catalog path when the host does not expand <catalog-root>.
  3. Review every active finding at its exact file and line.
  4. Separate confirmed behavior, ambiguous behavior, and false positives.
  5. If a false positive is accepted, suppress only its exact fingerprint, rule ID, and rule version with a concrete reason.
  6. Re-run the scan and report both active and suppressed counts.
  7. Stop before installation or permission grants unless the user separately authorized them.
  8. When evaluating the scanner itself, scan agent-skillguard/fixtures/positive (expected review findings) and agent-skillguard/fixtures/negative (expected clean result) with the bundled script, and inspect rules/non-coverage.json. The installed plugin package does not ship the upstream tools/verify_rule_corpus.py helper, so do not claim that helper ran; the two fixture scans are the supported local smoke test.

Packaging Notes

This catalog copy is intentionally self-contained: it includes the scanner, rule JSON, schemas, and public fixtures, but not the plugin's host metadata, examples, or large assets. The scanner resolves its rules relative to this skill directory, so invoke it through the bundled Python script rather than assuming a globally installed skillguard command.

Exit codes

  • 0: no active findings at or above the selected severity.
  • 1: at least one active finding requires review.
  • 2: the scan could not be completed reliably.

Read the full file on GitHub · 86 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago Changed 44a58ff54666
  2. 6d ago First seen · 86 lines · 55 tokens per session scan A f37391097023

Subscribe to this mod's changes

agent-skillguard is a skill published in the GitHub repository PracticalSwan/agent-skills (14 stars, last pushed 4d ago), licensed MIT. It adds 55 tokens to every session and 1,045 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-06.