Guardrails
341 mods. Block, rewrite or audit tool calls before they run.
-
subagent-effort
0.2.1
Run each subagent on any model (gpt-6, grok-4.7, ...) at any reasoning effort, through model and effort parameters on the Agent tool
HaoYan-A in Guardrails, updated
0
-
blast-radius
0.1.0
See what a risky command would change before it runs.
You in Guardrails, updated
0
-
blast-radius
0.1.0
See what a risky command would change before it runs.
You in Guardrails, updated
0
-
swamp
0.1.0
Gives Claude a tool that runs a swamp workflow and returns its result.
Mark Ellens in Guardrails, updated
0
-
slop-detector
0.1.0
Catches slop as Claude writes it: refuses a file edit that replaces code with a '... rest of the code' placeholder, and tracks TODO stubs, not-implemented throws, debug prints, @ts-ignore and other leftovers in a band above the prompt and a /slop pane.
slopshopper in Guardrails, updated
0
-
pin-everywhere
A pinned session keeps its Remote Control on (reconnected after app restarts) once machine-tag has put the machine icon in the title, so it shows by name and icon on every other computer and the phone; unpinning turns it back off
Tuncer Deniz in Guardrails, updated
0
-
blast-radius
0.2.0
Faengt gefaehrliche Shell-Befehle (Bash und PowerShell) ab, zeigt Befehl und Folge und fragt vor dem Ausfuehren nach; Standard aus, mit /ward fuer die Sitzung einschalten
Dan Enso in Guardrails, updated
0
-
image-peek
0.1.1
Previews pasted images and images Claude sends, with Open and Reveal buttons (Explorer on WSL/Windows, Finder on macOS)
Chanlito Hem in Guardrails, updated
0
-
context-saver
0.1.0
Refuses a main-agent Read over the token cap so a subagent reads the file and reports back.
MahadSalim in Guardrails, updated
0
-
lint-guard
0.1.0
An ordinary eslint wrapper: lints after every Write/Edit, a /lint command, and a band showing current findings.
interop-probe in Guardrails, updated
0
-
pr-proof
0.2.9
mod + skill
Shows the proof behind a PR under the reply that names it: checks, before/after screenshots, what was not checked, with Open and Reveal buttons
Chanlito Hem in Guardrails, updated
0
-
guardrails
1.1.0
Stops risky shell commands before they run: killing ports owned by Docker containers, printing secrets from the environment, rewriting your kubeconfig, and asks before kubectl writes on production contexts.
Lucas Leandro in Guardrails, updated
0
-
blast-radius
0.1.0
Before a shell command deletes or discards files, shows exactly what it would hit and waits for Cancel or Proceed.
JJ Englert in Guardrails, updated
0
-
greve-generale
0.2.0
Les outils de Claude se mettent en grève au hasard, pour des raisons très françaises
Greg in Guardrails, updated
0
-
done-gate
0.1.0
Claude cannot say "done" while the tests fail: a Stop gate runs your test command after every turn that changed files, and existing test files are locked so they cannot be patched green.
Daniil Bystrov in Guardrails, updated
0
-
block-destructive-commands
0.2.0
Denies Bash commands that match destructive patterns (recursive rm on roots, force push, hard reset, destructive SQL, disk formatting) before they run. A tool.call hook with a Bash matcher that returns { deny } without calling next.
claude-code-templates in Guardrails, updated
0
-
dense
0.1.1
A Claude Code plugin that folds everything per turn.
Jin Huang in Guardrails, updated
0
-
collision-guard
0.1.0
Предупреждает, если файл недавно менял другой чат Claude Code
Amir Aktay in Guardrails, updated
0
-
nixie-hold-promise
0.0.0
Spike: a tool.call hook that holds a call on a promise of its own
zgeoff in Guardrails, updated
0
-
publish-guard
1.0.0
Stops Claude from typing private terms into the browser or writing them into your publishing files, and keeps a log of what it blocked
Build Alone in Guardrails, updated
0
-
toyon-tally
0.1.0
Counts tool calls and adds a /tally command: the smallest mod, for the Claude integration test
toyon-dev in Guardrails, updated
0
-
buddy
0.11.1
Buddy: a pet sidekick for Claude Code. Sees all your chats, keeps your to-do list when Claude needs you, covers your secrets while you film, runs a team of helper agents and background chats, manages your skills, warns before a cold-cache message, and asks before sends and paid calls
Josh Pocock in Guardrails, updated
0
-
tailscale
0.3.2
Lets Claude query and modify the tailnet through the Tailscale API, using TS_API_KEY.
anderson-spider in Guardrails, updated
0
-
approve-blocked
0.1.0
Retries an auto-mode-blocked call up to three times so Claude Code's own permission prompt (and Moshi) can ask you.
adammhal in Guardrails, updated
0
-
compact-tail
0.1.1
Automatic compaction summarizes only the older part of the conversation and keeps the latest part word for word; a reopen tool brings back anything summarized, from a local archive
Tarl Raney in Guardrails, updated
0
-
ops-rules
0.1.0
Rules for the session started with LIFE_ROLE=ops: it decides and does not measure (analysis commands need the person's OK).
38kta-lab in Guardrails, updated
0
-
soundtrack
0.1.0
A soft soundtrack for Claude Code: a rising chord when Claude starts, a quiet tick while tools run, a chime when it's done.
RenegadeRocks in Guardrails, updated
0
-
pkexec-guard
0.2.0
Routes `sudo` in Bash calls to a password prompt the person can answer: pkexec on a Linux desktop, an osascript administrator dialog on macOS, and Windows sudo's UAC prompt on Windows.
Josh Kerr in Guardrails, updated
0
-
cortex-guard
0.1.2
Two rules no settings hook enforces, refused at the tool call: wiki pages and their generated ADR mirrors are written only through the Cortex wiki tool, and every git worktree lives at <repo>/.claude/worktrees/<name>/ (one created there is registered with disk_hygiene.py).
cdeust in Guardrails, updated
0
-
zetetic-autopilot
0.1.1
Automatic effort and model: every request runs at the effort its grade calls for (zetetic-genius) and the tool loop steps down the ladder; subagents run on the family their type and the turn's class call for (haiku, sonnet, opus, fable); under quota or context pressure everything goes one notch d...
cdeust in Guardrails, updated
0