Skip to content
Back to skills

Requirements

ASecurity

WHEN a user gives a concrete request without naming a skill, THE SYSTEM SHALL route the request through `/autoplan` to the matching skill, lifecycle chain, or orchestrator fallback, SO THAT users do not need to manually choose from the full skill catalog.

  • 47 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 1, 2026
ai-agentsgosecurity

Works with

  • cli

Security analysis

A100/100

Pro scans all 21 files and shows the line behind each finding

Scanned September 25, 2026

npx -y skills add rjmurillo/ai-agents --skill requirements --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Requirements?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Requirements
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rjmurillo-requirements/badge)](https://www.skillsdirectory.com/skills/rjmurillo-requirements)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
type: requirement
id: REQ-019
title: Autoplan Router Skill
status: implemented
priority: P2
category: developer-experience
epic: skill-routing
related:
  - TASK-019
  - REQ-039
issues:
  - 2828
  - 5385
created: 2026-07-03
updated: 2026-09-25
author: richard
---

# REQ-019: Autoplan Router Skill

## Requirement Statement

WHEN a user gives a concrete request without naming a skill,
THE SYSTEM SHALL route the request through `/autoplan` to the matching skill, lifecycle chain, or orchestrator fallback,
SO THAT users do not need to manually choose from the full skill catalog.

## Context

Issue #2828 requested a lazy router inspired by gstack autoplan. The implemented scope is the mechanical router skill and its Copilot CLI mirror. The open design question about whether `/autoplan` should retire the `orchestrator` agent is not part of this requirement.

## Acceptance Criteria

- [x] REQ-019-AC1: WHEN `/autoplan` is invoked, THE SYSTEM SHALL classify the request by intent family and size tier SO THAT routing starts from the user's outcome.
- [x] REQ-019-AC2: WHEN a request maps to a high-traffic workflow, THE SYSTEM SHALL route through the table in `.claude/skills/autoplan/SKILL.md` SO THAT common work avoids the orchestrator fallback.
- [x] REQ-019-AC3: WHEN a new capability is requested, THE SYSTEM SHALL run `programming-advisor` prior-art discovery before `/spec`, and SHALL invoke `buy-vs-build-framework` only for a strategic build, buy, partner, or defer decision, SO THAT repository governance is preserved without routing every narrow prior-art question through the strategic framework (REQ-039 criterion 6, DESIGN-037).
- [x] REQ-019-AC4: WHEN security triage is needed, THE SYSTEM SHALL route detection through `security-detection` and vulnerability review through `security-review` or `security-scan` SO THAT security paths use the canonical skills.
- [x] REQ-019-AC5: WHEN a decision is mechanical, taste-based, or sovereignty-bound, THE SYSTEM SHALL apply the matching handling rule SO THAT user-owned decisions are not auto-decided.
- [x] REQ-019-AC6: WHEN no routing row matches, THE SYSTEM SHALL apply the REQ-039 precedence: the long-tail resolver for a single-domain request, the lifecycle chain for a feature, bug, or shipping request, and `Task(subagent_type="orchestrator")` only for multi-domain, cross-cutting, or multi-agent work SO THAT long-tail work reaches the smallest owner.
- [x] REQ-019-AC7: WHEN the skill is published to Copilot CLI, THE SYSTEM SHALL include the generated `src/copilot-cli/skills/autoplan/SKILL.md` mirror SO THAT both harnesses expose the same router.
- [x] REQ-019-AC8: WHEN plugin source changes, THE SYSTEM SHALL bump both project-toolkit plugin manifests to the same patch version SO THAT release validation can prove parity.

## Out of Scope

- Retiring or deleting the `orchestrator` agent.
- Rewriting the root skill-routing guidance to delegate to `/autoplan`.

Files in this skill

  • EARS-TEMPLATE.md10 KB
  • README.md3.1 KB
  • REQ-001-pr-comment-handling.md1.7 KB
  • REQ-002-pr-comment-triage.md1.8 KB
  • REQ-003-multi-tool-artifact-build.md36.9 KB
  • REQ-004-agent-eval-harness-spike.md20.3 KB
  • REQ-005-command-skill-bundling.md11.2 KB
  • REQ-006-em-en-dash-enforcement.md15.3 KB
  • REQ-007-skill-catalog-prune-m1.md4.9 KB
  • REQ-008-review-axes-convergence.md26.4 KB
  • REQ-009-orphan-ref-validator.md9.7 KB
  • REQ-010-m4-rework-warning-recalibration.md7 KB
  • REQ-011-agent-skill-classification-audit.md6 KB
  • REQ-012-retro-fixes-pr-1965.md23 KB
  • REQ-013-m5-bot-cascade-pre-push.md14 KB
  • REQ-014-pr-description-extension-boundary.md4.3 KB
  • REQ-015-pr-iteration-cost.md13 KB
  • REQ-016-spec-step0-first-principles-gate.md20.6 KB
  • REQ-017-spec-memory-first-gate.md22.2 KB
  • REQ-018-skill-catalog-prune-m2-session-cluster.md3.2 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.