$ loading_
Check installed community skills for updates with diffs and explicit approval.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "auto-updater" skill from askskill: 1. Download https://raw.githubusercontent.com/anthropics/claude-for-legal/main/legal-builder-hub/skills/auto-updater/SKILL.md 2. Save it as ~/.claude/skills/auto-updater/SKILL.md 3. Reload skills and tell me it's ready
Check whether my installed community skills have updates. Show the diff first and do not apply anything automatically.
A list of updatable skills, diffs between current and latest versions, and a prompt for user approval before any update is applied.
Update my skills. If there are changes, show me the diff for each one and ask for confirmation.
It checks upstream versions for installed skills, shows the changes for each item, and applies updates only after explicit approval.
Anything new for my installed skills? Tell me which ones have updates and highlight key risk points.
It returns the list of skills with updates and highlights risk areas such as hooks, MCP config, permission scope, or purpose changes.
Developers or DevOps users can review full diffs and risk signals before upgrading installed skills, helping avoid unreviewed code or permission changes entering their environment.
When a user wants to know whether installed skills have newer versions, this skill checks each source version and summarizes available updates for manual review.
If an update changes hooks, .mcp.json, tool permissions, or stated purpose, this skill highlights those differences so users can review them more carefully.
The README describes a skill that loads the installed skill list and update preferences, checks each skill source for a newer commit SHA, and compares the pinned version with the latest one. When updates exist, it presents full diffs and highlights risky changes such as hooks, MCP config, expanded tool access, network calls, command execution, or purpose changes. It applies nothing without explicit approval and recommends re-running security checks on new versions.
~/.claude/plugins/config/claude-for-legal/legal-builder-hub/CLAUDE.md → installed skills + auto-update prefs.Community skills improve. This skill notices when, shows you what changed, and applies updates only with your explicit approval.
Installed skills are code running inside your privileged legal environment. An upstream repository can be compromised, transferred to a new owner, or simply change behavior in ways you don't want. This skill is designed so that no update is ever applied without you reading the diff and approving it. That's not a preference — it's the design.
~/.claude/plugins/config/claude-for-legal/legal-builder-hub/CLAUDE.md → installed skills (with version/commit SHA), update preferences (notify / manual).
For each skill in the installed list:
For each update, show the full diff:
# [skill-name] — [installed SHA] → [latest SHA]
## SKILL.md changes
[unified diff]
## hooks/hooks.json changes
[unified diff — FLAG: hooks can execute arbitrary code]
## .mcp.json changes
[unified diff — FLAG: MCP servers run with your credentials]
## Other files
[list of added/removed/modified files with diffs]
Then run the trust check:
hooks/hooks.json change? Hooks can execute arbitrary shell commands. Show the diff prominently and ask the user to confirm they understand what the new hooks do..mcp.json change? New or changed MCP servers can access your environment. Same treatment.allowed-tools or tools frontmatter expand? New tool access is a permission escalation.description or stated purpose change? A skill that claimed to "review NDAs" and now claims to "send contracts" has repurposed itself.Re-run the full skills-qa scan against the NEW version before applying the
update. A skill that was clean at v1.0 can ship a poisoned v1.1 — the
GlassWorm pattern (a trusted publisher, an established skill, a minor
version bump that carries the payload). Install-time trust does not
transfer to updates.
Rules:
skills-qa Step 1.5 category — refuse
the update by default and explain why. Emit the new-version REFUSE
output verbatim.hooks/hooks.json, .mcp.json, allowed-tools/tools
frontmatter, new Bash/WebFetch/WebSearch access, new external URLs,
new file-write paths outside the skill directory, or the description
frontmatter FORCES a human-approval prompt and cannot be bypassed by a
clean LLM scan. The scan is a signal; the human is the gate.restrictive allowlist mode, the read-only subagent
is MANDATORY here — do not apply an update in restrictive mode without…
It checks whether installed community skills have updates, shows the diffs, and applies updates only after explicit user approval.
No. The provided material explicitly says no update is applied until the user reviews the diff and approves it.
The docs say it prominently flags changes to hooks/hooks.json, .mcp.json, expanded tool permissions, and risks like new network calls, writes outside the skill directory, command execution, or purpose changes. See the source repository for more details.
Search watched skill registries, preview matches, and inspect SKILL.md before installing.
Draft DMCA notices, triage incoming claims, and prepare counter-notice drafts.
Standardize new matter intake and generate matter files plus a structured log.
Compare proposed handbook edits and identify cross-reference and state supplement impacts.
Review offer letters and restrictive covenants with jurisdiction-specific hiring checks.
Manage matter workspaces to separate, switch, and archive client contexts.
Safely review and install community skills with explicit user approval.
Review skill PRs with severity-rated feedback on quality and repository conventions.
Suggests relevant community skills based on your task and recent plugin activity.
Safely uninstall, disable, and re-enable community skills with audit logging.
Safely uninstall hub-installed community skills with confirmation and audit logging.
Create, refine, validate, and restructure AgentSkills and SKILL.md files.