advisor-guard
by sipeng · ★ 0 · updated 3 days ago
Under a Fable or Mythos main loop, tell the model never to consult the advisor tool, so the session does not pay for a second model.
Install
/plugin marketplace add SipengXie2024/claude-mods/plugin install advisor-guard@sipeng-modsThen run /reload-plugins or start a new session. Requires Claude Code 2.1.287+.
What it hooks into
From an automated scan with claude plugin validate, not a security review. Mods run with your permissions; read the code before installing.
- Events
prompt.attachmentprompt.compose- API calls
$.session.model- Reach
- level 0 of 3
Origin
- How it got here
- Automated scan of public GitHub repos, via the CC0 dataset awesome-claude-code-mods (scan of ).
- Repository
- github.com/SipengXie2024/claude-mods
- Path in repo
mods/advisor-guard- Source commit
f4f872cthe exact code the scan read- Author
- sipeng @SipengXie2024
- License
- No license declared — ask the author before reusing code.
- First published
- Last pushed
- First listed here
Name and description come from the mod's own manifest. We link to the code; we don't host it, and it stays under its author's license.
More from SipengXie2024/claude-mods
context-memory
SipengXie2024/claude-mods
Lets the working model keep a task note, archives the conversation at compaction, carries the latest note verbatim into the new context, and provides scoped history search and read tools. Makes no model requests of its own.
jev
SipengXie2024/claude-mods
Jev effort routing for Opus, Mythos, GPT-6 Astra and GPT-6.1 Sol (policy.v15): start effort per prompt from per-model kind tables that start low and hold a raised level while the task goes on, judged by OpenAI's Decisions API over the conversation as Claude Code holds it (user messages and final replies, newest last, in one request with the follow-up fault question), and a raise carried into...
kernel-model-hint
SipengXie2024/claude-mods
Under Fable 5.2 or Mythos 5.2, rewrite the system-prompt note on their kernel-optimization training limits so the model suggests Mythos 5.1 for accelerator kernel work.
magpie-tasks
SipengXie2024/claude-mods
Long-running task windows for claude-magpie sessions: an architect per long task, subagent model and effort routing by rules judged through Jev, a task registry and log, worktrees with integration, acceptance and diagnosis of daily tasks, Telegram delivery, startup recovery, rating invitations and calibration reminders. Active only in sessions that talk to the Magpie gateway.
math-render
SipengXie2024/claude-mods
Draws the TeX math in assistant replies as Unicode in the terminal: inline formulas on one line, display formulas as character art.
mode
SipengXie2024/claude-mods
Per-repo general/academic mode: /mode writes .claude/settings.local.json; paper repos get a one-time hint.
no-bg-poll
SipengXie2024/claude-mods
Block sleep-and-read polling of background tasks; they send a completion notification.
peer-send
SipengXie2024/claude-mods
Scenario driver of the magpie-tasks e2e scripts: a session loaded with it sends a text to another session, the way the Router hands a Telegram message to an architect.
reply-language
SipengXie2024/claude-mods
Restate the settings' reply language beside every prompt, so long English context does not pull replies into English.
subagent-opus
SipengXie2024/claude-mods
When the main loop runs on a Fable or Mythos model, start Agent-tool subagents on Opus instead of inheriting it.
tg-router
SipengXie2024/claude-mods
Telegram Router: dispatches each request to a background worker, routes replies to that worker, posts its answers, and shows a progress card per run; under claude-magpie it leaves worker models, long tasks, decisions and scores to magpie-tasks. Load with --plugin-dir in the claude-telegram session only.
workflow-opus-guard
SipengXie2024/claude-mods
Under a Fable or Mythos main loop, deny Workflow scripts whose agent() calls do not pin model: 'opus'.