Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsBlogPro
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges
  • Chrome Extension
  • Skill Manager

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Rpi Plan

ASecurity

Create a cited phased implementation specification with behavioral oracles, explicit scope and measurable automated and manual acceptance criteria.

5 stars
0 votes
0 copies
0 views
Added 10/4/2026
ai-agents

Security Analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned 10/4/2026

$npx -y skills add juan294/cc-rpi --skill rpi-plan --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Rpi Plan?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Rpi Plan
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/juan294-rpi-plan-cc-rpi/badge)](https://www.skillsdirectory.com/skills/juan294-rpi-plan-cc-rpi)

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

Download with Pro
Files
SKILL.md
---
name: rpi-plan
description: "Create a cited phased implementation specification with behavioral oracles, explicit scope and measurable automated and manual acceptance criteria."
---

Create an implementation plan for: the request

Process:
1. Read controlling instructions/contracts and directly mentioned files completely,
   including accepted research or assessment when present. Revalidate their baseline against the
   actual worktree; cite valid findings and decisions rather than reconstructing them.
2. Find relevant code, patterns and docs. Keep narrow work with the parent. Useful
   independent assignments stay within planning and state objective, permitted
   read-only actions/files, evidence/output, resource limits and completion condition.
   Use available slots and inspect every required result and coverage gap.
3. Inspect discovered implementation to the depth required by unresolved decisions;
   reuse valid prior reads instead of rereading every discovered file.
4. Present your understanding and ask only for material information the request,
   accepted artifacts and repository cannot resolve. Do not repeat supplied decisions.
5. Resolve remaining uncertainties with targeted evidence when needed.
6. Present design options with trade-offs.
7. Propose phase structure, get feedback.
8. Write detailed plan with separate phase files.
9. Read [the pseudocode notation](references/pseudocode-notation.md) and use it
   for nontrivial changes. Ground acceptance in executable or checkable evidence.
10. When the plan specifies behavior, prefer pointing at an executable or checkable
    artifact (a failing test, a module with the semantics to match, a mockup, a rubric)
    over describing the behavior in prose.
11. Separate automated vs. manual success criteria.
12. Write a **Stuck states and recovery** section: for every state the change
    can enter that fails closed, blocks work, or degrades output (a barrier,
    lock, fallback, disabled path, cached error), state who sees it and what
    they see; how it ends without insider knowledge (automatic recovery, or a
    visible prompt that starts the fix); and the test that proves it ends or
    that the user is told. A safety test alone is not enough; each stuck state
    also needs a recovery-or-disclosure test. "Recovery requires a manual step"
    is acceptable only when the user is shown that step.
13. Write a **Consumer sweep** section when the change alters what a shared
    function, type, cache format or fixture returns or stores (especially on
    failure): list every caller and writer found by search, with the command
    used, and mark each covered by a phase or explicitly excluded with a
    reason. Include test fixtures, E2E helpers and scripts that write the
    same data.
14. Identify independent work units within each phase: no file overlap or
    dependency on another unit's output. Mark those units `[batch-eligible]`
    for local worktree coordination, with one integration owner and no working
    branch push/PR creation. Keep phase execution and acceptance sequential.
15. Use at most three temporary [NEEDS CLARIFICATION] markers and resolve
    every one before final acceptance. Do not fill unknowns with placeholders.
16. Resolve every material decision before plan acceptance. A missing required
    investigation or reviewer result remains an explicit acceptance gap.
17. Apply the [durable handoff](references/handoff.md) contract in the plan/notes,
    naming accepted decisions, evidence limits and next-phase entry conditions.
    Revalidate actual worktree state on resume.

Save to docs/plans/YYYY-MM-DD-[description].md
Phase files: docs/plans/YYYY-MM-DD-[description]-phases/phase-N.md

No unresolved questions in the final plan. Present the saved plan and phase-file
paths, acceptance criteria and remaining decisions, then stop at the planning
boundary. Do not implement unless the user explicitly authorized that next workflow.

## Execution and acceptance

Use the scope and authorization already supplied in the request. Resolve routine
implementation choices from repository evidence. Complete authorized local work,
review, repair and applicable verification before its acceptance gate. An explicit
instruction can authorize continuation across phases; otherwise stop at the stated
phase boundary. Production, publication, destructive actions and new scope retain
their actual authorization requirements. Preserve durable artifacts before cleanup.

Attribution

juan294juan294
View sourceSee grades on GitHubMore from juan294 →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

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 (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".

1100021 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

698461 votes

Writing Skills

Create and manage Claude Code skills in HASH repository following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns), UserPromptSubmit hook, and the 500-line rule. Includes validation and debugging with SKILL_DEBUG. Examples include rust-error-stack, cargo-dep...

3931 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3421 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Amp, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Grok Build, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

741 votes
View all in ai-agents →