Run writing-great-skills over every SKILL.md to cut verbosity

[TBarregren]

Summary

Every skill in this repo has been written organically and the SKILL.md files have grown wordy. Run Matt Pocock's writing-great-skills over each skill and rewrite it to be as lean as possible. If a skill can be cut down substantially without losing behaviour, that is a win.

Motivation

SKILL.md content is loaded into the agent's context when the skill triggers, so tokens spent on redundant prose are a direct, recurring cost paid on every invocation. Leaner skills mean faster, cheaper, more reliable triggering and easier maintenance.

Scope

Apply writing-great-skills to each of the following (current size for reference):

Skill	Words	Lines
skills/orchestrate/SKILL.md	~11350	334
skills/release/SKILL.md	~1760	237
skills/coder/SKILL.md	~1290	76
skills/coding-standard/SKILL.md	~1146	92
skills/init/SKILL.md	~1135	131
skills/doctor/SKILL.md	~820	74
skills/push/SKILL.md	~784	75
skills/commit/SKILL.md	~741	68

orchestrate is the clear outlier and the highest-value target.

Approach

• Run writing-great-skills against one SKILL.md at a time.
• Prefer trimming redundancy, repetition, and over-explanation; move genuinely-needed detail into on-demand agents.d/ / reference files rather than the always-loaded body where it makes sense.
• One skill per commit/PR so each rewrite is reviewable in isolation.

Acceptance criteria

• Every SKILL.md above has been passed through writing-great-skills and rewritten.
• No behavioural regression: triggering conditions, guardrails, and workflow steps are all preserved.
• Existing tests still pass (audit.py + pytest); the JS engine's structural tests still hold.
• Measurable reduction in word count where the content allows, especially orchestrate.