How do I install the Conductor skill?

Run npx skillmds@latest add zedarvates/conductor in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Conductor safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Conductor?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Conductor free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Conductor?

zedarvates (@zedarvates) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Conductor by zedarvates — Agent Skill

Name: Conductor
Author: zedarvates

Conductor

Route a high-level goal to an ordered, local-first plan of capabilities — the generalised router — and optionally EXECUTE the plan's read-only steps. Reads the capability registry/curator, composes steps ordered by the system's layers (SENSE→DECIDE→ACT→REMEMBER→GOVERN→DEPLOY), annotates each local vs cloud with a concrete command, and estimates the goal's effort. The executor runs the safe analysis steps unattended and confirm-gates anything that mutates state or escalates to the cloud. Use when the user states a goal and you need to decide which botte capabilities to use in what order (or run them), or asks "how should I approach X with this toolkit".

zedarvates 9d28034 6 files · 20.9 KB Updated Sep 17, 2026

File contents

conductor — goal → ordered plan of capabilities

The router, generalised: not "which model tier?" but "given this goal, which capabilities, in which order, and what stays local?". The plan is the product — you (or the agent) execute it; the Conductor never runs anything itself.

python -m skills.conductor.cli "test my desktop app and report crashes"
python -m skills.conductor.cli "reduce token cost and deploy on my project" --json

# Execute the plan (read-only steps run; mutating/cloud steps are gated):
python -m skills.conductor.cli "audit my project and report metrics" --execute
python -m skills.conductor.cli "..." --execute --dry-run      # preview, runs nothing
python -m skills.conductor.cli "..." --execute --confirm      # also run gated steps

How it plans

Curate — picks the capabilities relevant to the goal ([[capabilities]] curator, local lexical match, 0 tokens).
Order — sorts them by the system's layers (SENSE → DECIDE → ACT → REMEMBER → GOVERN → DEPLOY): understand → decide → act → remember.
Annotate — each step gets a concrete command, a local/cloud flag, and the reason it's there.
Estimate — the goal's effort tier ([[auto_router]]) tells you whether any step's reasoning will escalate to the cloud.

Output: an ordered list of steps, 0 cloud tokens to produce. It composes the module collection into a coherent plan per goal — the conductor of the system.

Executing the plan

The plan can be run, not just read. The executor classifies every step:

safe — read-only analysis (directives_audit, metrics, infra_advisor, checkup, cluster, llm_backends) → runs unattended, 0 cloud tokens.
gated — mutates state, generates artifacts, or escalates to the cloud → runs only with --confirm / confirm=true.
needs_args — the command still has an unfilled <placeholder> → never runs.

--dry-run classifies everything and runs nothing (a preview). A failing step yields a non-zero exit so CI can react. The runner is injectable, so the behaviour is fully unit-tested without spawning subprocesses.

Exposed via [[llm_mcp]] as conduct (plan) and execute_plan (plan + run safe steps). Built on [[capabilities]], [[auto_router]]; pairs with the [[control_loop]] (measure savings → adapt the routing).

---name: conductordescription: Route a high-level goal to an ordered, local-first plan of capabilities — the generalised router — and optionally EXECUTE the plan's read-only steps. Reads the capability registry/curator, composes steps ordered by the system's layers (SENSE→DECIDE→ACT→REMEMBER→GOVERN→DEPLOY), annotates each local vs cloud with a concrete command, and estimates the goal's effort. The executor runs the safe analysis steps unattended and confirm-gates anything that mutates state or escalates to the cloud. Use when the user states a goal and you need to decide which botte capabilities to use in what order (or run them), or asks "how should I approach X with this toolkit".---# conductor — goal → ordered plan of capabilitiesThe router, generalised: not "which model tier?" but "given this **goal**, whichcapabilities, in which order, and what stays local?". The plan is the product —you (or the agent) execute it; the Conductor never runs anything itself.```bashpython -m skills.conductor.cli "test my desktop app and report crashes"python -m skills.conductor.cli "reduce token cost and deploy on my project" --json# Execute the plan (read-only steps run; mutating/cloud steps are gated):python -m skills.conductor.cli "audit my project and report metrics" --executepython -m skills.conductor.cli "..." --execute --dry-run # preview, runs nothingpython -m skills.conductor.cli "..." --execute --confirm # also run gated steps```## How it plans1. **Curate** — picks the capabilities relevant to the goal ([[capabilities]] curator, local lexical match, 0 tokens).2. **Order** — sorts them by the system's layers (SENSE → DECIDE → ACT → REMEMBER → GOVERN → DEPLOY): understand → decide → act → remember.3. **Annotate** — each step gets a concrete command, a local/cloud flag, and the reason it's there.4. **Estimate** — the goal's effort tier ([[auto_router]]) tells you whether any step's reasoning will escalate to the cloud.Output: an ordered list of steps, **0 cloud tokens** to produce. It composes themodule collection into a coherent plan per goal — the conductor of the system.## Executing the planThe plan can be *run*, not just read. The executor classifies every step:- **safe** — read-only analysis (`directives_audit`, `metrics`, `infra_advisor`, `checkup`, `cluster`, `llm_backends`) → runs unattended, 0 cloud tokens.- **gated** — mutates state, generates artifacts, or escalates to the cloud → runs only with `--confirm` / `confirm=true`.- **needs_args** — the command still has an unfilled `<placeholder>` → never runs.`--dry-run` classifies everything and runs nothing (a preview). A failing stepyields a non-zero exit so CI can react. The runner is injectable, so thebehaviour is fully unit-tested without spawning subprocesses.Exposed via [[llm_mcp]] as `conduct` (plan) and `execute_plan` (plan + run safesteps). Built on [[capabilities]], [[auto_router]]; pairs with the [[control_loop]](measure savings → adapt the routing).

Frequently asked questions