Skip to content
Back to skills

Update Skill

ASecurity

Update an existing skill or knowledge entry in phucbm/skills. Use when the user says "update my X skill", "update X knowledge", or "add this change to my X skill/knowledge".

  • 2 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added October 13, 2026
ai-agentsgoshellbashgit

Works with

  • cli

Security analysis

A100/100

Scanned October 13, 2026

npx -y skills add phucbm/skills --skill update-skill --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Update Skill?

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

Security grade badge for Update Skill
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/phucbm-update-skill/badge)](https://www.skillsdirectory.com/skills/phucbm-update-skill)

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
---
name: update-skill
description: Update an existing skill or knowledge entry in phucbm/skills. Use when the user says "update my X skill", "update X knowledge", or "add this change to my X skill/knowledge".
allowed-tools: Bash Read Write Edit
---

## Structure

### Skills — flat
- `skills/<topic>/SKILL.md` — thin wrapper. Update only if trigger wording, description, or knowledge pointers change.

### Knowledge — namespaced, shareable
- `knowledge/<category>/<slug>.md` — the actual content. This is usually what gets updated.
- A knowledge file may be referenced by multiple skills — check before editing. If shared, note which other skills will be affected.

## Steps

0. **Read the prompt** — if arguments were passed with the slash command (e.g. `/update-skill add the new retry logic to my groq skill`), treat them as the user's description of what to update and proceed to step 1. If no arguments were given, ask: "Which skill or knowledge do you want to update, and what should change?" and wait for the answer before continuing.

1. **Pull latest**:
   ```shell
   gh repo clone phucbm/skills /tmp/phucbm-skills 2>/dev/null || git -C /tmp/phucbm-skills pull
   ```

2. **Surface related files** — scan `skills/` and `knowledge/` for anything related to the topic. List all candidates and flag shared knowledge files:
   > "Files that may need updating:
   > - `skills/groq/SKILL.md` (trigger description)
   > - `knowledge/groq/streaming-integration.md` (actual content — also referenced by skills/X)
   > Should I update both, or just the knowledge file?"

3. **Show a diff for each file that will change**:
   ```
   FILE:     knowledge/groq/streaming-integration.md
   CURRENT:  <existing section>
   PROPOSED: <updated section>
   REASON:   <why>
   ```

4. **Wait for user confirmation** before applying anything.

5. **Apply edits**, bump patch version in `.claude-plugin/plugin.json`, commit and push:
   ```shell
   git -C /tmp/phucbm-skills add .
   git -C /tmp/phucbm-skills commit -m "knowledge: update <topic>"
   git -C /tmp/phucbm-skills push
   ```

6. Confirm the push URL to the user.

## Rules
- Follow Anthropic's official skill writing best practices:
  https://github.com/anthropics/skills/blob/main/skills/skill-creator/SKILL.md
  Key points: `description` is the primary trigger — keep it specific and slightly pushy; prefer surgical edits to stay under 500 lines.
- Never silently overwrite — always show the diff first
- If README.md description for this topic changed, update that row too
- Prefer surgical edits over full rewrites
- `SKILL.md` rarely needs updating — only if frontmatter `description`/`when_to_use` or knowledge pointers change
- Knowledge file refs must use clickable markdown links:
  - Single: `Read [\`@/knowledge/<category>/<slug>.md\`](/knowledge/<category>/<slug>.md) and apply...`
  - Multiple: `Read and apply:` followed by a bullet list of `[@/knowledge/...](...)` links
- Knowledge files are generic — examples referencing specific projects are fine, but never add project-specific data as facts

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.