Using NecturaLabs Skills
Overview
Initializes the NecturaLabs skill suite. This skill runs at conversation start and after every agent change to ensure all NecturaLabs skills are active and properly configured.
Initialization Flow
digraph init {
"Conversation starts / agent changes" [shape=doublecircle];
"Already initialized?" [shape=diamond];
"Superpowers loaded?" [shape=diamond];
"Load superpowers:using-superpowers" [shape=box];
"Check superpowers installed" [shape=box];
"Superpowers installed?" [shape=diamond];
"Prompt user to install" [shape=box];
"User installs?" [shape=diamond];
"Block NecturaLabs skills" [shape=box];
"Load agent-context-loader" [shape=box];
"Ready (reviews active for session)" [shape=doublecircle];
"Conversation starts / agent changes" -> "Already initialized?";
"Already initialized?" -> "Ready (reviews active for session)" [label="yes, fully aware"];
"Already initialized?" -> "Superpowers loaded?" [label="no"];
"Superpowers loaded?" -> "Check superpowers installed" [label="yes"];
"Superpowers loaded?" -> "Load superpowers:using-superpowers" [label="no"];
"Load superpowers:using-superpowers" -> "Check superpowers installed";
"Check superpowers installed" -> "Superpowers installed?";
"Superpowers installed?" -> "Load agent-context-loader" [label="yes"];
"Superpowers installed?" -> "Prompt user to install" [label="no"];
"Prompt user to install" -> "User installs?";
"User installs?" -> "Load agent-context-loader" [label="yes"];
"User installs?" -> "Block NecturaLabs skills" [label="no"];
"Load agent-context-loader" -> "Ready (reviews active for session)";
}
Step 1: Ensure Superpowers Is Loaded
If superpowers:using-superpowers has not been invoked yet in this session, invoke it first. NecturaLabs skills build on top of superpowers and require its session initialization to be complete before proceeding.
Step 2: Verify Superpowers Dependency
NecturaLabs skills require superpowers to be installed. Check that the superpowers:requesting-code-review skill is available (the old superpowers:code-reviewer agent no longer exists — do not probe for it).
If NOT installed, tell the user:
NecturaLabs skills require the superpowers plugin. Install it with:
/plugin marketplace add obra/superpowers
/plugin install superpowers@superpowers-dev
Do not allow necturalabs:iterative-code-review or necturalabs:iterative-security-audit to run without superpowers installed. Other NecturaLabs skills may run independently.
Step 3: Load Context
Invoke necturalabs:agent-context-loader to load global CLAUDE.md and project AGENTS.md.
Ongoing: Mandatory Post-Change Reviews
Decision Flow — After ANY Code Changes
digraph review_trigger {
"Agent finishes code changes" [shape=doublecircle];
"Changes security-related?" [shape=diamond];
"Invoke necturalabs:iterative-security-audit" [shape=box];
"Invoke necturalabs:iterative-code-review" [shape=box];
"Both loops complete" [shape=box];
"Show summary" [shape=doublecircle];
"Agent finishes code changes" -> "Changes security-related?";
"Changes security-related?" -> "Invoke necturalabs:iterative-security-audit" [label="yes"];
"Changes security-related?" -> "Invoke necturalabs:iterative-code-review" [label="no"];
"Invoke necturalabs:iterative-security-audit" -> "Invoke necturalabs:iterative-code-review" [label="audit chains into review"];
"Invoke necturalabs:iterative-code-review" -> "Both loops complete";
"Both loops complete" -> "Show summary";
}
When to invoke necturalabs:iterative-code-review
After ANY code changes — feature implementation, bug fixes, refactoring, test additions, config changes. No exceptions.
When to invoke necturalabs:iterative-security-audit (BEFORE code review)
When changes are security-related (see the skill's description for the full trigger list). Security audit takes priority — it runs first and chains into code review automatically.
Available Skills
| Skill | Purpose | When to invoke |
|---|---|---|
necturalabs:iterative-code-review |
Industry-standard code review loop | After any code changes, before commit/merge |
necturalabs:iterative-security-audit |
OWASP/CWE security audit loop | When changes touch security-sensitive code |
necturalabs:agent-context-loader |
Loads CLAUDE.md + AGENTS.md into context | On init and after context switches |
necturalabs:agents-md-manager |
Creates/updates project AGENTS.md | Manual (/agents-md-manager) or after plan execution |
necturalabs:git-workflow |
Conventional Commits + worktree isolation | When committing or starting multi-commit work |
necturalabs:update-plugins |
Updates all plugin marketplaces and plugins concurrently | Manual (/update-plugins) |
necturalabs:docs-manager |
Creates/maintains project docs/ with ADRs, design docs, guides | Manual (/docs-manager) or when user asks to document |
necturalabs:test-manager |
Classifies and triages a project's tests, routes to the level specialists | Test work with no obvious level, spanning levels, or suite-wide |
necturalabs:unit-test-manager |
Unit tests — one unit, one process, no I/O | Writing, updating, or fixing unit tests |
necturalabs:integration-test-manager |
Tests across a process boundary — DB, HTTP, queue, filesystem | Writing, updating, or fixing integration tests |
necturalabs:e2e-test-manager |
End-to-end and browser tests of critical journeys | Writing, updating, or fixing E2E tests |
necturalabs:comment-manager |
Comment and doc-comment rules across languages, with a per-language matrix | Writing or changing code that carries comments |
necturalabs:using-necturalabs |
This skill — initializes everything | On init and after agent handoffs |
Skill Priority When User Asks
If NecturaLabs skills are installed and the user asks for "code review", "security audit", any work on tests, or any work on comments, docstrings or in-source documentation, invoke the NecturaLabs skill — not the default behavior or any other plugin's version. NecturaLabs skills are complementary to superpowers and layer additional standards on top.
Re-Initialization
Re-run this initialization when:
- A new conversation begins
- A subagent is dispatched or returns
- The main agent context is switched or compressed
- The user explicitly asks to reload skills
Skip re-initialization ONLY if the current agent is already fully aware of all NecturaLabs skills and their mandatory review requirements are understood.