Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
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

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Reference

ASecurity

Point at a reference — treat a provided example as a map, not an answer, and prove comprehension before implementing. Use when the user says "use this as a reference", "make it like this", or supplies example code, a mockup, or a competitor's UX to build from.

2 stars
0 votes
0 copies
0 views
Added 9/28/2026
ai-agentsgoreact

Security Analysis

A100/100

Scanned 9/28/2026

Install to Claude Code

$npx -y skills add ajitta/know-your-unknowns --skill reference --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Reference?

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

Security grade badge for Reference
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/ajitta-reference/badge)](https://www.skillsdirectory.com/skills/ajitta-reference)

More formats (shields.io, HTML) on the badges page.

Files
SKILL.md
---
name: reference
description: >
  Point at a reference — treat a provided example as a map, not an answer, and prove
  comprehension before implementing. Use when the user says "use this as a reference",
  "make it like this", or
  supplies example code, a mockup, or a competitor's UX to build from.
argument-hint: "<reference file/path/description>"
---

# Reference — A Reference Is Another Map

Best way to give a model a map: **give it another map**.
Origin: Point at a Reference — see skills/loop/references/talk-source.md.

## Iron Rule

**Reference is not an answer to copy verbatim — it is material for understanding intent and behavior.**
Before implementing, **prove comprehension** with a semantics map.

## Procedure

1. Read reference to the end; summarize **intent, core behavior, invariants**.
2. Present 4-category analysis:
   - **Behaviors to preserve** — the reference's reason to exist
   - **Parts to transform for current environment** — language/framework/scale differences
   - **Unnecessary or dangerous parts** — must not be brought into this project
   - **Parts improvable beyond the reference** — where we can do better
3. For port/transform work, also present **proof of comprehension (semantics map)**:
   key reference excerpts ↔ new-environment counterparts side by side, gotcha points
   where behavior subtly differs, and an edge-case table with a **Match** column —
   identical / equivalent / changed ("equivalent" = same decision, different surface).
   Number every note and row so a correction can name one, and give each row the
   reference location it came from (`file:line`, section, or timestamp) so the user can
   check the reading against the source.
4. Contrast with current project conventions (code style, dependencies, test
   approach); present an **application plan**.
5. **Sign-off gate — nothing is implemented until the user signs off.** They reply
   `semantics confirmed`, or correct any row by its number ("note 5", "budget
   exhaustion row") and the map is revised before any code. In-session, offer the same
   choice once with AskUserQuestion: implement now / revise the map / stop.
6. On sign-off, **port the reference's existing tests first**, then implement. Record
   intentional deviations from the reference per the **notes** skill.

## Output

Artifact tool if available → else `.unknowns/<YYYY-MM-DD>-reference-<slug>.html` → else
markdown; always echo the assembled reply in chat.
Numbered mapping rows (reference excerpt ↔ counterpart plan) and edge-case rows each
carry **approve / request-change**; selections assemble into a `semantics confirmed`
reply or a numbered correction template.
Details: skills/loop/references/output-routing.md

## What Can Be a Reference

Existing implementation code, same algorithm in another language, working HTML
mockup, competitor product UX, past project design docs, test code,
screenshots/video, actual example of desired output. E.g., an HTML mockup as the
map when building a React component.

Attribution

ajittaajitta
View sourceMore from ajitta →
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

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1074701 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', ...

695601 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.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, 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.

691 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →