Skip to content
SkillModel-led2,719 tokens in use

Leader

Turns a one-line idea into a goal brief that an agent can run alone for hours, with named off-limits, proof by command output and a check afterwards.

22,429 stars on the repository MIT LicenseKKKKhazix/khazix-skills · leader
/plugin marketplace add KKKKhazix/khazix-skills /plugin install khazix-skills@khazix-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Leader costs

160tokens idle
2,719tokens in use
6,075tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter8/15
  • Structure16/25
  • Examples0/15
  • Output0/15
  • Size13/15
  • Resources15/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity80
  • Maintenance84
  • Community60
  • Maturity28
  • Momentum40

Leader casts you as the person with the idea, the agent that uses the skill as the manager, and a goal-mode agent as the executor. The manager investigates your codebase, asks at most five questions that only you can answer, and writes a brief of up to 4000 characters for the /goal command.

The brief is built for long unattended runs. It fixes what the executor may touch, how success is proved, and which shortcuts count as cheating. After the run, the manager re-checks the result and reports in five lines.

What You Get

  • A single goal brief of up to 4000 characters, ready to paste after /goal
  • A list of decisions the skill made for you, with the default chosen and the cost of a wrong guess
  • Acceptance checks written as commands, with anti-cheat rules and stop conditions
  • A short verification report after the run, based on checks you never saw

Common Use Cases

  • Handing an agent a project to finish unattended overnight
  • Writing a brief for a test-fixing or refactoring job without allowing deleted tests
  • Setting up an exploratory research goal with capped, sourced conclusions
  • Splitting a large job into parallel briefs with clear boundaries

What Makes It Unique

The brief focuses on what must not be done, because a goal with no limits invites shortcuts such as skipping tests or mocking the code under test. Each rule traces back to something measured or something you decided. The manager keeps a few hidden spot checks for the final verification, the executor cannot grade itself, and the brief has a progress file, a blocked file and a three-failure rule to stop wasted effort.

Under The Hood

The skill is a single instruction file plus two reference documents, with no scripts. It tells the agent to explore your repository and run the real build, test and lint commands to measure baselines, and to search the web for industry facts when it can. It asks you questions in chat and returns the brief as text in the reply. The brief it writes tells the executor to keep a progress file and a blocked file in the project, and the skill itself keeps its hidden checks in a scratch area on the manager side.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers