guardloop

guardloop is a command for Claude Code from samibs/skillfoundry. It costs 0 tokens per session (1,658 once invoked), scanned A, original, MIT.

A command-line guardrail tool that tracks repeated failure patterns in coding sessions and turns recurring problems into project rules.

In plain words
What is it for?
Use it to review pattern counts, scan a codebase for known failure patterns, check hook status, and promote proven patterns into rule files.
Why use it?
It helps catch mistakes that happen repeatedly and makes lessons from past coding work available in future sessions.

Command for Claude Code

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.

agentmods
npx agentmods add commands/samibs/skillfoundry/guardloop
Clone the repo
git clone --depth 1 https://github.com/samibs/skillfoundry

Made for: Claude Code.

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 guardloop

README.md
[![agentmods](https://agentmods.dev/badge/commands/samibs/skillfoundry/guardloop.svg)](https://agentmods.dev/commands/samibs/skillfoundry/guardloop)
Your own site
<a href="https://agentmods.dev/commands/samibs/skillfoundry/guardloop"><img src="https://agentmods.dev/badge/commands/samibs/skillfoundry/guardloop.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,658 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00000 $0.01658
Opus 5 $0.00000 $0.00829
Sonnet 5 $0.00000 $0.00332
Haiku 4.5 $0.00000 $0.00166

Measured yesterday against content hash ef29233b35d6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

guardloop 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 yesterday.

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.

.claude/commands/guardloop.md · 220 lines

How it starts

The opening of the file, as written. The whole thing — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/guardloop — Adaptive Learning Guardrail Engine

Analyzes recurring failure patterns harvested from coding sessions and promotes them into enforced rules in agents/_guardloop-rules.md.

Powered by GuardLoop (github.com/samibs/guardloop.dev) × SkillFoundry.


Usage

/guardloop              Pattern frequency report (default)
/guardloop analyze      Same as above — full report with candidates
/guardloop promote      Promote ready patterns → agents/_guardloop-rules.md
/guardloop scan         Scan codebase for known failure patterns right now
/guardloop status       Show pattern counts + hook health
/guardloop reset        Reset all pattern counters (use after major cleanup)

Instructions

You are the GuardLoop Engine — the self-learning layer that converts observed LLM failures into enforced guardrails. You learn from this project's real history, not from theoretical rules.


Default / analyze — Pattern Frequency Report

Step 1: Run analysis script

bash scripts/guardloop-analyze.sh

Step 2: Read the pattern state directly for additional context

.claude/hooks/state/guardloop-patterns.json

Step 3: Read last 5 entries tagged guardloop from the knowledge base to show recent examples

memory_bank/knowledge/errors-universal.jsonl

(filter lines where "tags" array contains "guardloop", take the last 5)

Step 4: Present the report:

GuardLoop Analysis — <date>
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  Patterns tracked:    10
  Total detections:    N
  Ready to promote:    N
  Already promoted:    N

  [table from script output]

  Recent detections:
    [last 3 from knowledge base]

If there are patterns ready to promote, recommend: Run /guardloop promote to generate guardrails.


promote — Promote Patterns to Agents

Step 1: Run promotion script

bash scripts/guardloop-promote.sh

Step 2: Read the updated agents/_guardloop-rules.md to confirm the new rules

Read the full file on GitHub · 220 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. yesterday First seen · 220 lines · 0 tokens per session scan A ef29233b35d6

Subscribe to this mod's changes

guardloop is a command published in the GitHub repository samibs/skillfoundry (12 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,658 tokens. 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 2026-09-03.