ainova-systems
OrganizationIntelligence pack for software engineering based on Ainova Systems best practices
Categories
Indexed Skills (39)
dev-review-changes
Reviews pending changes against the project's rules and reports findings with a severity verdict. Read-only - never edits, stages, or commits.
spec-approve
Records the owner's approval so a reviewed spec enters the autonomous queue. Autonomous mode only - under supervision a plan with no open question is already approved.
spec-audit-docs
Audits documentation claims against the code and classifies each mismatch as drift or violation. Read-only by default.
dev-handoff
Writes a self-contained prompt a fresh session pastes to continue this work, for when context runs short or a session ends.
git-create-release
Cuts a release - pending-step review, owner gate, version, changelog, tag - per the project's release policy. Owner-invoked only; release timing is not the model's call.
git-finalize-pr
Drives an open pull request to merge-ready - CI green, every review thread answered, one outcome label. Opening it is git-open-pr; merging is git-merge-pr.
git-merge-pr
Merges an accepted pull request behind guard checks, then syncs the base branch and cleans up. Owner-invoked only - merge timing is not the model's call.
git-open-pr
Opens a pull request for the current branch against its target, filling the repo's template. Driving it to green afterwards is git-finalize-pr.
git-review-pr-comments
Triages review comments on a pull request - fix, discuss, or decline with a reason - and leaves no thread unanswered.
spec-cancel
Retires a spec as cancelled or superseded with the reason recorded, and reconciles the docs it had drafted.
spec-continue
Resumes an in-progress spec: audits inherited work for drift, then continues from the first unticked work step.
spec-create
Turns a stated task or captured brief into a spec's requirements when no tracker item covers it, then hands to spec-plan. Tracker intake is spec-pull.
spec-decision
Records an architecture decision behind a three-condition gate. A standing law belongs in its rule, behavior in the feature doc - not here.
spec-document
Writes or updates one documentation artifact - feature, rule, glossary, model, or architecture - from what the code actually does.
spec-execute-next
Picks the highest-value approved spec, drives it to an outcome-labeled pull request, and resets the workspace. Autonomous mode only.
spec-execute
Executes a planned spec through parallel subagents, one testable slice each, gates re-run until dry. Refuses while any question on the spec stays open.
spec-init
Scaffolds the in-repo docs substrate for spec-driven work and migrates existing documentation into it. Runs once, before the first spec.
spec-plan
Writes the plan for a spec whose requirements already exist - coverage table, phases, checkboxed work steps. Both intakes hand off here.
spec-pull
Pulls one tracker item into a spec, read-only against the tracker, and updates that same spec in place on every re-run. With no tracker item, spec-create is the path.
spec-validate
Fact-checks a planned spec against the repository before execution and turns each gap into an open question that may re-block. Read-only on code.
dev-run-tests
Runs typecheck, lint and tests scoped to what changed, then analyses the failures. The gate every commit passes first.
git-commit-push
Commits pending work as one verified milestone and pushes it. Stops at the push - opening the pull request is git-open-pr.
git-scan-secrets
Scans a diff, the working tree, or branch history for credentials before they reach a remote, and classifies every hit.
spec-answer
Resolves a blocked spec's open questions with the developer and records what each answer changed. Questions owned by someone outside the session stay open.
spec-close
Confirms a merged spec actually shipped - PR merged, steps ticked, docs reconciled - then archives it per the project's convention.
git-resolve-conflicts
Resolves merge or rebase conflicts by what each side intended, never by picking a hunk, then re-runs the full gates.
intelligence-compact-context
Reduce rules, agents, and skills without changing behavior or teaching terse output
intelligence-review-skills
Audit the intelligence layer for duplication, drift, size, hardcoded paths and framing
intelligence-update
Interpret an update plan and verify breaking post-conditions
intelligence-add-agent
Create new specialized agent
intelligence-add-rule
Create new intelligence rule
intelligence-add-skill
Create new skill
intelligence-extract-skill
Extract observed workflow into a reusable skill
intelligence-learn-from-context
Capture one approved lesson from a session in an established Intelligence project
intelligence-learn-from-repository
Recover and complete first-time Intelligence repository onboarding
dev-run-roadmap
Implement one roadmap batch through a merge-ready PR
intelligence-install-adapter
Research, implement, and enable a tool adapter
intelligence-sync
Sync intelligence to enabled adapters
intelligence-uninstall-adapter
Disable an adapter and assess its generated output
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.