conjure-agent

conjure-agent is a command for coding agents from john-farina/claude-plugins. It costs 27 tokens per session (4,221 once invoked), scanned B, original, MIT.

A command for creating or updating a Claude Code agent: a separate worker that can handle a task with its own context, model, or tool access.

In plain words
What is it for?
Use it to define agents for independent work, parallel tasks, or jobs that need sustained specialist context.
Why use it?
It helps decide when a separate worker is useful and avoids creating one when an ordinary command would be simpler.

Command

Part of the conjure plugin — 7 commands shipped together

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/john-farina/claude-plugins/conjure-agent
Clone the repo
git clone --depth 1 https://github.com/john-farina/claude-plugins

Or install conjure, the plugin that ships this one along with the rest of its 7 commands.

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 conjure-agent

README.md
[![agentmods](https://agentmods.dev/badge/commands/john-farina/claude-plugins/conjure-agent.svg)](https://agentmods.dev/commands/john-farina/claude-plugins/conjure-agent)
Your own site
<a href="https://agentmods.dev/commands/john-farina/claude-plugins/conjure-agent"><img src="https://agentmods.dev/badge/commands/john-farina/claude-plugins/conjure-agent.svg" alt="Measured on agentmods" height="20"></a>
Per session 27 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,221 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. 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 $0.00027 $0.04221
Opus 5 $0.00014 $0.02110
Sonnet 5 $0.00005 $0.00844
Haiku 4.5 $0.00003 $0.00422

Measured 4d ago against content hash ffe4636cdb37, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

conjure-agent scanned grade B with 1 finding 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.

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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

cat ~/.claude/conjure-config.md 2>/dev/null
plugins/conjure/commands/conjure-agent.md · 399 lines

How it starts

The opening of the file, as written. The whole thing — 399 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Craft Agent

Creates or updates a Claude Code agent definition (~/.claude/agents/<name>.md or project-local).


Phase 0 — "Should this be an agent at all?"

Before creating anything, apply this gate. An agent is the RIGHT tool only when one of these is true:

  1. Context pollution — prior subtask results (1000+ tokens) would degrade reasoning on the next step
  2. True parallelism — independent workstreams that gain nothing from sharing a context window
  3. Tool overload — the job needs 15+ tools, or fundamentally conflicting behavioral modes (e.g., "never edit" and "always edit")
  4. Sustained specialization — the agent accumulates domain context across many calls that a generalist can't maintain

If none apply: a well-prompted single Claude call (or a skill/command) achieves equivalent results at 10-15x lower token cost. Say so and stop.

Redirect table — say this and stop if matched:

Request sounds like… Actually is… Say…
"agent to commit my changes" Skill — user-triggered, inline workflow "This is better as a skill — run /conjure-skill <name>"
"agent to run lint whenever I save" Hook — fires automatically "This is better as a hook — run /conjure-hook <description>"
"agent to explain architecture" Skill (reference) "This is better as a skill — run /conjure-skill <name>"
"agent to search the codebase" Agent ✅ — isolated context, specific tools Continue
"agent to investigate crashes" Agent ✅ — spawned programmatically Continue
"agent to always do X" Hook — automatic, event-driven "This is better as a hook — run /conjure-hook <description>"

If updating an existing agent: skip to Phase 1 — the decision was already made.


Phase 0.5: Load User Preferences

Before doing anything else, read both conjure config files if they exist:

cat ~/.claude/conjure-config.md 2>/dev/null
cat .claude/conjure-config.md 2>/dev/null

Follow all instructions found in ## agents and ## global sections throughout every phase of this command. Project-local instructions (.claude/conjure-config.md) take precedence over global ones when they conflict. If neither file exists, proceed with defaults.

Read the full file on GitHub · 399 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 First seen · 399 lines · 27 tokens per session scan B ffe4636cdb37

Subscribe to this mod's changes

conjure-agent is a command published in the GitHub repository john-farina/claude-plugins (1 stars, last pushed 3mo ago), licensed MIT. It adds 27 tokens to every session and 4,221 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other commands, from other repositories