mod plugins

by claude-code-templates

Organization control mod: withholds the http and process nouns from $ so no plugin beneath it can reach the network or spawn processes, refuses user-tier plugins by name allowlist or by the $ calls their source declares, and withholds the Bash tool (shellPolicy "deny", default) or, in "guardrail" mode, denies well-known network clients as a bypassable speed bump. Hooks engine.create, plugin.register and tool.call; seat it in the managed prepend tier.

by claude-code-templates

/aitmpl opens a component browser above the prompt: the aitmpl.com catalog (agents, commands, MCPs, settings, hooks, skills, loops, mods) read live from the site's public JSON through $.http.fetch. Pick a type or a trending component, filter, page, open a component and install it from there ($.process.run with the fixed CLI argv and a validated path, no shell), drop its install command into the prompt, or open its page in your browser. Hooks session.start, command.run and ui.render on AbovePrompt; browsing costs no tokens.

by claude-code-templates

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.

by claude-code-templates

Screens every prompt going into the conversation and every reply coming out of it with TypeSafe's Jev, a System One decision model, the way TypeSafe's guardrails cookbook does: one request per message asks a battery of hazard questions (jailbreak, harm or a crime, a diagnosis or a dosage, self-harm) and scores how much harm complying would do. The thresholds live here, in a named policy (strict or permissive), and turn the probabilities into pass, review (the user is asked), block (the prompt is dropped, or the reply withheld) or support. Falls back to the engine's built-in classifier when no key is set.

by claude-code-templates

Picks the model and the reasoning effort for each task with TypeSafe's Jev, a System One decision model reached either through TypeSafe's own API or the Vercel AI Gateway. Routes subagent models at agent.spawn and the main loop's effort and model at turn.step, in both directions, falling back to the engine's built-in classifier when no key is set.

by claude-code-templates

Keeps the skill listing out of the context window and lets TypeSafe's Jev, a System One decision model, pick at most one skill per prompt from the skills' descriptions, the way TypeSafe's skill-suggestion cookbook does: one request ranks every skill and asks whether the prompt needs a skill at all, a second re-reads the top three with their SKILL.md and can reject them all. The winner's SKILL.md is attached to the prompt, so the skills can be hidden from the model altogether (/jev-skill-suggestion:setup makes them user-invocable-only and /context counts them at 0). Falls back to the engine's built-in classifier when no key is set.

by claude-code-templates

Asks the user, in the engine's AskUserQuestion dialog ($.ui.ask), before Claude edits or overwrites a file larger than a configurable line count; denies by default when nobody can answer (headless runs). A tool.call hook on Edit and Write using $.fs.read.

by claude-code-templates

Rewrites npm and npx invocations to pnpm (or yarn / bun) before the Bash tool runs them and tells the model what ran through context. A tool.call hook that forwards a modified event (the modifying placement).

by claude-code-templates

Denies Edit, Write and NotebookEdit calls on sensitive paths (.env files, lockfiles, CI workflows, git internals, private keys) with a glob allowlist override. A tool.call hook with an array matcher.

by claude-code-templates

Redacts API keys, tokens, JWTs, private keys and connection strings from every tool result before the model reads them, notes the redaction through context, and refuses to echo a redacted placeholder back into a Bash command. A tool.call hook in the after placement.

by claude-code-templates

Times every tool call and wraps the engine's ToolUse rendering with a colored duration badge from the surface's own element table, on the terminal, Desktop and mobile. Hooks tool.call and ui.render using JSX.

by claude-code-templates

One hook on * that records every event on $ (tool calls, prompts, turns, other plugins' fs/http/process calls) as JSON lines with origin plugin and tier, duration and outcome, including denials. Buffered and flushed to a JSONL file through $.fs at the end of every turn.

by claude-code-templates

Short-circuits repeated WebFetch calls for the same URL and prompt within a session, answering { result } from cache without a network round trip. A tool.call hook with TTL and size limits.

by claude-code-templates

Overrides the built-in WebSearch tool and routes queries to the Exa search API through $.http.fetch, answering in WebSearch's own result shape, with automatic fallback to the built-in search. A tool.call hook that replaces a core tool.

by Fernando Moraes

A live checklist above the prompt of the conversation's loose ends: questions left unanswered, points made in passing, caveats and deferred work it moved past, kept up to date by a small model after each turn. Needs function hooks (early access) and an interactive terminal.