# Plan Feature

> A Claude skill from Karthick-Ramachandran/persist-os.

- **Type:** Skill
- **Install:** `agentstack add skill-karthick-ramachandran-persist-os-plan-feature`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [Karthick-Ramachandran](https://agentstack.voostack.com/s/karthick-ramachandran)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [Karthick-Ramachandran](https://github.com/Karthick-Ramachandran)
- **Source:** https://github.com/Karthick-Ramachandran/persist-os/tree/main/.claude/skills/plan-feature
- **Website:** https://persist-os.pages.dev

## Install

```sh
agentstack add skill-karthick-ramachandran-persist-os-plan-feature
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Skill: Plan Feature

## Purpose

Turn product intent into a scoped engineering plan that an implementation agent can follow safely.

## Inputs

- Feature PRD.
- Acceptance criteria.
- Relevant module or architecture docs.
- Known constraints or release target.

## Required Reading

- `docs/10-architecture/ARCHITECTURE.md`
- `docs/10-architecture/FILE_WRITE_POLICY.md`
- `docs/20-security/SECURITY_MODEL.md`
- `docs/50-quality/QUALITY_GATES.md`
- `docs/60-engineering/ENGINEERING_STANDARDS.md`
- Relevant feature docs under `docs/40-features/`
- Relevant module docs under `docs/30-modules/`
- Relevant ADRs under `docs/adrs/`

## Output Files

- `docs/40-features//PLAN.md`
- `docs/40-features//TASKS.md`
- `docs/40-features//ARCHITECTURE_IMPACT.md`
- `docs/40-features//TEST_PLAN.md`

## Process

1. Restate the feature objective and acceptance criteria.
2. Identify modules, docs, templates, and tests affected.
3. Document architecture impact and ADR needs.
4. Break work into ordered tasks with clear completion evidence.
5. Define tests from requirements, risk, security invariants, and regressions.
6. For module requests, treat the module as a mini product and create feature delivery docs before
   implementation tasks.

## Stop Conditions

Stop and request human decision if:

- Requirements are missing or contradictory.
- Architecture impact cannot be determined.
- The plan conflicts with engineering standards.
- A task requires changing accepted non-goals.
- A module request tries to start implementation before PRD, acceptance, architecture impact, test
  plan, and tasks exist.

## Quality Bar

- Tasks are ordered and independently reviewable.
- Tests map to acceptance criteria and risks.
- Architecture impact is explicit.
- Engineering standards are accounted for in tasks and completion evidence.
- The plan does not include implementation code when only planning is requested.

## Source & license

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

- **Author:** [Karthick-Ramachandran](https://github.com/Karthick-Ramachandran)
- **Source:** [Karthick-Ramachandran/persist-os](https://github.com/Karthick-Ramachandran/persist-os)
- **License:** MIT
- **Homepage:** https://persist-os.pages.dev

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

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-karthick-ramachandran-persist-os-plan-feature
- Seller: https://agentstack.voostack.com/s/karthick-ramachandran
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.