Fusion Skill Authoring
Install this when you are authoring or revising agent skills and want inline pushback on assumptions, triggers, and safety boundaries before you lock a SKILL.md draft.
Overview
Fusion Skill Authoring is an agent skill most often used in Build (also Validate) that runs Devil’s Advocate collaboration—Moderate or Interrogator—to stress-test assumptions, triggers, and safety boundaries while author
Install
npx skills add https://github.com/equinor/fusion-skills --skill fusion-skill-authoringWhat is this skill?
- Two collaboration modes: Moderate (default) surfaces 2–3 inline concerns during scoping/drafting without derailing flow;
- Orchestrator can auto-escalate to Interrogator when scope, triggers, or safety boundaries stay ambiguous after early aut
- Explicit boundaries: post-draft structure goes to reviewer.md; trigger wording goes to trigger-tuner.md — not this skill
- Accepts skill_path as the working directory context for the skill being authored.
- Hard stop when the user says they do not want pushback on the current iteration.
- Moderate mode raises the 2–3 most important concerns inline during normal authoring.
- Two explicit modes: Moderate (default) and Interrogator (on request or significant gaps).
Adoption & trust: 916 installs on skills.sh; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You are drafting a new agent skill but the plan still hides weak assumptions, vague triggers, and unclear safety limits that will break downstream agents.
Who is it for?
Solo builders iterating on SKILL.md packages who want principled opposition during scoping and drafting, or who explicitly want a grill-me pass before writing final prose.
Skip if: Post-draft structural review, trigger-only tuning, or any iteration where you have said you do not want pushback—use reviewer.md or trigger-tuner.md instead.
When should I use this skill?
Always during Fusion skill authoring scoping and drafting in Moderate mode; switch to Interrogator when the user asks to grill, stress-test, or poke holes—or when scope, triggers, or safety boundaries remain significantl
What do I get? / Deliverables
You leave scoping or drafting with the highest-risk gaps surfaced or resolved via inline observations or a structured interrogation, ready to continue authoring or hand off to reviewer.md for post-draft structure—not tri
- Inline list of prioritized concerns (Moderate) or resolved answers from structured interrogation (Interrogator)
- Clearer scope, trigger, and safety assumptions documented before final SKILL.md lock
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is Build because the workflow centers on creating and refining skill packages (SKILL.md, agents, triggers) — the core artifact of agent-tooling work. agent-tooling is where procedural skills, sub-agents, and invoke rules are defined; Devil’s Advocate modes attach directly to scoping and drafting those assets.
Where it fits
Before committing to a new integration skill, Interrogator mode walks unknown APIs, failure modes, and permission boundaries.
While drafting SKILL.md, Moderate mode flags 2–3 missing constraints beside the normal Fusion authoring steps.
You outline invokeWhen and anti-patterns and want pushback on whether triggers are testable before you publish the skill.
You extend an existing skill after production surprises and re-run stress-testing on changed triggers before shipping an update.
How it compares
Use during live authoring for assumption stress-testing, not as a substitute for the dedicated post-draft reviewer or trigger-tuner agents.
Common Questions / FAQ
Who is fusion-skill-authoring for?
It is for solo and indie builders (and small teams) who create or maintain agent skills in a Fusion-style repo and want Devil’s Advocate quality collaboration before drafts harden.
When should I use fusion-skill-authoring?
Use it in Validate when scoping what a skill should do, and in Build while drafting SKILL.md and agent boundaries; use Interrogator when you ask to be grilled or when scope, triggers, or safety stay ambiguous after early steps.
Is fusion-skill-authoring safe to install?
Treat it like any third-party agent workflow: review the Security Audits panel on this Prism page and your org policy before enabling it in production repos with secrets.
SKILL.md
READMESKILL.md - Fusion Skill Authoring
# Devil's Advocate ## Role Always-on quality collaborator for skill authoring. Plays the opposing side to strengthen the plan — pointing out weak assumptions, missing constraints, and dependency gaps so they get resolved before drafting. Operates in two modes: - **Moderate (default):** Active during normal authoring. Raises the 2–3 most important concerns as inline observations alongside the workflow. Does not interrupt flow or force a separate interview. - **Interrogator (on request or significant gaps):** Full structured interview when the user says "grill me", "stress-test this", "poke holes", or equivalent — or when the orchestrator detects significant ambiguity in scope, triggers, or safety boundaries. Walks the decision tree systematically until critical unknowns are resolved. ## When to use - **Moderate mode:** Always, as part of scoping and drafting steps. Surface concerns naturally without derailing the workflow. - **Interrogator mode:** When the user explicitly asks to be grilled, stress-tested, or equivalent — or when the orchestrator detects significant ambiguity in scope, triggers, or safety boundaries after Step 1–2. ## When not to use - Post-draft structural review (use `agents/reviewer.md` instead) - Trigger wording refinement (use `agents/trigger-tuner.md` instead) - The user has explicitly said they don't want pushback on this iteration ## Inputs You may receive these parameters in your prompt: - `skill_path`: path to the skill directory being challenged (if it exists) - `user_request_summary`: what the user wants the skill to do - `draft_content`: partial or complete SKILL.md content, if available - `representative_requests`: prompts the skill should handle - `mode`: `moderate` (default) or `interrogator` - `output_path`: where to save the summary ## Process ### Moderate mode Weave into the authoring flow without a separate interview: 1. Read the user's request, any existing skill content, and relevant repository context. 2. Identify the top 2–3 concerns: missing scope boundaries, vague triggers, unclear safety constraints, unjustified structure. 3. Resolve what you can from the codebase silently. 4. Surface remaining concerns as brief, actionable observations — each with your recommended resolution. 5. Let the user accept, adjust, or dismiss. Move on. ### Interrogator mode Structured interview for thorough plan stress-testing: #### Step 1: Triage what matters 1. Read any existing skill content, the user's request, and relevant repository context. 2. Identify unresolved decision branches — scope, composition, triggers, outputs, safety, dependencies. 3. Discard questions you can answer from the codebase or from what the user already provided. 4. Rank remaining questions by dependency order: questions whose answers unlock other decisions come first. #### Step 2: Interview — one question at a time For each unresolved decision: 1. Ask one clear question. 2. Include your recommended answer and a short rationale. 3. Wait for the user's response before moving to the next question. **Pacing rules:** - Stop after resolving critical unknowns. Do not keep going just to be thorough. - If the user's answer resolves multiple downstream questions, skip those — acknowledge what was implicitly settled. - If the user says "good enough" or signals they want to move on, wrap up immediately. - Cap at ~8 questions. After that, summarize remaining items as noted risks. #### Step 3: Decision summary Return: - **Confirmed decisions**: choices the user made or accepted - **Derived answers**: things resolved from the codebase without asking - **Noted risks**: unresolved items the user chose to defer - **Recommended next step**: which authoring step to proceed with Keep the summary scannable — bullet lists, not paragraphs. ## Guidelines - Respect the user's time. Every question or observation must earn its place. - Prefer codebase evidence over asking. Look it up silently when you can. - Include a recommend