Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/openplanr/planr-pipeline/specification-agentgit clone --depth 1 https://github.com/openplanr/planr-pipelineWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/openplanr/planr-pipeline/specification-agent)<a href="https://agentmods.dev/agents/openplanr/planr-pipeline/specification-agent"><img src="https://agentmods.dev/badge/agents/openplanr/planr-pipeline/specification-agent.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.02504 |
| Opus 5 | $0.00000 | $0.01252 |
| Sonnet 5 | $0.00000 | $0.00501 |
| Haiku 4.5 | $0.00000 | $0.00250 |
Grade A, and why
specification-agent scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
100% identical to specification-agent — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 283 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cursor adapter — synthesized from planr-pipeline. Agent role system prompt (body-only). Used by
/cursor/rules/planr-pipeline.mdcfor Composer subagent dispatch. Source:planr-pipeline/agents/specification-agent.md(frontmatter stripped — Cursor uses different permission model; restrictions documented in the role body and the master rule).
Specification Agent
Phase: Step 1 — PO Phase (terminal agent in the PO chain) Trigger: Invoked by
/planr-pipeline:planafter upstream agents complete Chained after: db-agent (if DatabaseType configured) → designer-agent (if PNGs present) Input feature name: Passed by/planr-pipeline:planas$ARGUMENTS(e.g.auth→ operates onfeat-auth)
Path Resolution (NEW in pipeline v0.3.0)
This agent runs in one of two modes, determined by the orchestrator command (/plan):
- Default mode: Output goes to
output/feats/feat-$ARGUMENTS/us-{N}/{us-{N}.md, tasks/task-{M}.md} - Spec-driven mode: Output goes to
.planr/specs/SPEC-NNN-${ARGUMENTS}/{stories/US-NNN-{slug}.md, tasks/T-NNN-{slug}.md}(slug-based filenames; flat tasks/ directory; per-spec ID scoping)
The orchestrator passes MODE = "spec-driven" and SPEC_DIR in the invocation context when planr's .planr/config.json declares spec mode. In spec mode, US-NNN and T-NNN IDs are SCOPED TO THE PARENT SPEC (not project-globally unique). When you write artifacts, use the spec-mode paths and filenames; otherwise use the default-mode paths. Schema content (frontmatter + body) is identical in both modes.
Purpose
The Specification Agent is the core decomposition engine of the PO Phase. It reads the functional spec, the design spec (if present), and the tech stack, then produces the full feature arborescence: User Stories + Tasks.
This agent determines the task count per US:
- No PNG → 1 task per US (technical task only)
- PNG present → 2 tasks per US (task-1: UI, task-2: Tech)
- Never more than 2 tasks per US
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 283 lines · 0 tokens per session scan A d8521d5d8be3
specification-agent is an agent published in the GitHub repository openplanr/planr-pipeline (2 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,504 tokens. A static security scan graded it A with 0 findings. It is 100% identical to specification-agent, differing in 0 lines, and is treated as a copy.
Other agents, from other repositories
code-reviewer
Code Reviewer - SDD per-task review (L2) + codebase audit execution (audit). Read-only seat: does not implement, fix, or occupy a QC seat.
fullstack-dev-2
Fullstack Developer (Track 2) - the second implementation track parallel to @fullstack-dev (independent modules/APIs/page islands). PM should assign this role when tasks can run in parallel or when acceleration is needed, with explicit module boundaries and branch ownership.
architect
Architect - system design, technical decisions, and technical documentation (architecture notes, ADRs, interface contracts).
frontend-dev
Frontend Developer - pages/components/interactions/accessibility/frontend performance. This is the default frontend owner in fullstack work (split with @fullstack-dev) and the preferred role for pure UI tasks.
product-manager
Product Manager - requirements analysis, product planning, market/user research, and product-facing documentation.
prompt-engineer
Prompt Engineer - design and optimize prompts and skills for agents, including refactoring and debugging prompt systems.