Nova Explore

Shared Execution Policy

This file is the supporting behavioral contract for /nova-plugin:explore and the deprecated /nova-plugin:nova-explore compatibility entrypoint. Prefer the direct command; the compatibility name remains only for the current major-version migration window.

• Resolve natural-language and explicit KEY=value inputs using ../_shared/parameter-resolution.md; explicit non-conflicting values take precedence.
• Apply ../_shared/safety-preflight.md before side effects. Never infer approval, destructive scope, credentials, or output destinations.
• Follow ../_shared/output-contracts.md and ../_shared/artifact-policy.md; report completed, skipped, and blocked validation truthfully.
• Respect the frontmatter tool boundary. Missing inputs, unavailable dependencies, overlapping user changes, or repository-policy conflicts are blockers rather than permission to broaden scope.

Execution

1. Parse $ARGUMENTS against the workflow-specific inputs below.
2. Read only the context required for the requested scope.
3. Apply the workflow contract and its strict output format.
4. Stop before unauthorized side effects; otherwise validate in proportion to risk and report residual risk.

Workflow Contract

> Generated from workflow-specs/behaviors.v2.json. This block is authoritative. Run node scripts/generate-behavior-surfaces.mjs --write after changing the IR; if explanatory text below conflicts, fail closed.

Generated Behavior Index

• Purpose: Align understanding and identify unknowns or risks without proposing solutions.
• Canonical inputs: INPUT(required aliases=REQUEST,CONTEXT); PERSPECTIVE(optional aliases=MODE default="observer" exact="observer","reviewer"); DEPTH(optional default="standard" exact="lite","standard","deep")
• Decision entries: 4; canonical routes and variants: explore {"PERSPECTIVE":"reviewer"}, explore {"PERSPECTIVE":"observer","DEPTH":"lite"}, explore {}, explore {"DEPTH":"deep"}.
• Workflow steps: resolve-input → route → analyze → emit
• Output: mode=chat; order=observations or what is clear → uncertainties or review questions → potential risks or risk signals; severity=none.
• Deviation/failure: mode=forbid; failure order=status → missing input → allowed values → safe next action.
• Full IR: runtime/contracts/explore.json#behaviorContract embeds the complete decision table, invariants, stops, field definitions, validation, and failure contract from the same source. Detailed guidance below may not override it.

Purpose

Quickly align understanding and identify unknowns/risks without proposing solutions.

Inputs

| Parameter | Required | Default | Notes | Example | | ------------- | -------- | ---------- | ------------------------------------ | ----------------- | | PERSPECTIVE | No | observer | observer or reviewer | reviewer | | INPUT | Yes | N/A | Requirement, diff, logs, design text | PR diff text... |

Outputs

• observer: Observations / Uncertainties / Potential risks.
• reviewer: What is clear / Review questions / Risk signals.
• Chat output only.

Workflow

1. Parse PERSPECTIVE.
2. Hub routing policy:

• observer -> nova-explore-lite
• reviewer -> nova-explore-review

1. Emit structured analysis output only.

Examples

• Natural trigger: Use explore to quickly align on this incident report.
• Explicit trigger: explore PERSPECTIVE=reviewer INPUT="Product requirement draft".

Safety

• Do not provide solutions, implementation plans, or code.
• Separate facts from assumptions.

Detailed Contract

QUICK EXPLORATION

You are Claude Code acting as a senior engineer / tech lead.

This command is for quick understanding and cognitive alignment, not for solving problems.

---

INPUT PARAMETERS

From $ARGUMENTS, extract the following:

PERSPECTIVE (Optional)

Choose the perspective for this exploration:

• observer (default) → Neutral observation and fact-gathering
• reviewer → Review mindset with critical questions

If not specified, use observer perspective.

PERSPECTIVE:

---

EXECUTION RULES

• Do NOT write code
• Do NOT propose solutions or designs
• Do NOT suggest refactors or optimizations

Focus only on:

• What is clearly understood
• What is uncertain or ambiguous
• Where risks may exist due to gaps in understanding

Keep the output concise and practical.

Language constraints:

• Avoid words like: "should", "recommend", "solution", "implement"
• Prefer: "observed", "suggests", "may indicate", "unclear", "appears"

---

OUTPUT FORMAT (STRICT)

Output format varies based on PERSPECTIVE:

If PERSPECTIVE = observer:

Observations

• Clearly stated facts from the input
• Direct, obvious inferences (clearly labeled)

Uncertainties

• Missing information
• Ambiguous behavior or intent
• Assumptions being made

Potential risks

• Risks caused by misunderstandings or unknowns
• No mitigation or suggestions

---

If PERSPECTIVE = reviewer:

What is clear

• Confirmed understanding based on provided input
• Explicitly separate facts from interpretations

Review questions

• Questions a reviewer would raise
• Focus on correctness, clarity, and assumptions
• Avoid hypothetical redesign questions

Risk signals

• Correctness risks
• Boundary or edge-case risks
• Operational or maintenance risks
• No mitigation or next steps

---

4.0 VARIANT PROFILES

• Default or PERSPECTIVE=observer: normal read-only exploration.
• PERSPECTIVE=reviewer: former explore-review behavior.
• DEPTH=lite: former explore-lite behavior.
• DEPTH=deep: former senior-explore behavior; EXPORT_PATH requires explicit artifact-write scope and the shared safety preflight.

Compatibility command presets are defaults only; explicit non-conflicting user inputs win. Runtime contracts remain the complete behavior source for each variant.

END OF COMMAND

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: lliangcol
• Source: lliangcol/llm-plugins-fusion
• License: MIT

Install and usage instructions live in the source repository linked above.