\r \r

Best Skill Recommendations\r

\r Primary role: Evaluate and recommend skills already discovered by an upstream skill search. Only when no upstream results exist does this skill independently search clawhub, then apply the same evaluation logic.\r \r

Prerequisites\r

\r Before this skill can operate, confirm the following:\r \r

• clawhub CLI is installed and available in the agent's PATH. This skill issues clawhub search, clawhub list, clawhub install, and clawhub uninstall commands. If the binary is absent, all discovery and install steps will fail.\r
• Authentication: Run clawhub login in advance. The clawhub CLI stores credentials in its own default config path (managed by the CLI, not by this skill). No additional environment variables are required by this skill.\r
• Permissions in scope: This skill will only:\r
  1. Read the installed skill list via clawhub list (read-only).\r
  2. Run clawhub install \x3Cskill> or clawhub uninstall \x3Cskill> — only after explicit user confirmation at the Pre-Install Gate (Step 5).\r It will not access other system files, credentials, or APIs beyond the clawhub CLI.\r \r

Mandatory Store Policy\r

\r

1. Prefer upstream results first. If any upstream skill search has already returned candidates, use those directly — do not re-search.\r
2. Only when no upstream candidates exist: search clawhub directly.\r
3. Never claim exclusivity; both public and private registries are valid.\r
4. Before install, always summarize source, version, and notable risk signals.\r
5. If a fresh search is needed, run clawhub search \x3Ckeywords> and report output before evaluating.\r \r

Auto-Trigger and Collaboration\r

\r

• Preferred entry: triggered after an upstream skill search has already produced a candidate list. Reuse those results directly.\r
• Standalone entry: triggered when the user asks to recommend/install/compare skills but no upstream results exist. In this case, independently search clawhub and produce candidates before evaluating.\r
• Never re-search if usable upstream candidates are already available.\r \r

Workflow\r

\r

0) Check for Upstream Candidates (first)\r

Before doing anything else:\r

• If an upstream skill search has already returned a candidate list → skip to Step 2 using those results.\r
• If no upstream candidates exist → proceed to Step 1.\r \r

1) Self-Discover (only when no upstream results)\r

Clarify the user's need:\r

• target task(s)\r
• priority (speed/stability/features/safety)\r
• constraints (region, cost, runtime)\r \r Then search via clawhub:\r

clawhub search \x3Ckeywords>\r
```\r
Return the command output and build a candidate list annotated with source and version.\r
\r
### 2) Evaluate Candidates\r
Present the candidate list (from upstream or self-discovered), annotated with source and version.\r
\r
### 3) Evaluate Installed Skills and Compatibility\r
Enumerate installed skills by running:\r
```\r
clawhub list\r
```\r
Parse the output to get the current installed skill set. Compare each installed skill against candidates:\r
- overlap: full / partial / complementary\r
- conflict risk: command/workflow collision, duplicated automation, behavior mismatch\r
- coexist feasibility: high / medium / low\r
\r
### 4) Recommend Replace vs Coexist\r
Per candidate, output one decision:\r
- Replace existing skill(s)\r
- Coexist with boundaries\r
- Do not install\r
\r
Include reasons and trade-offs.\r
\r
### 5) Pre-Install Gate (required)\r
Before any install action, present:\r
- source (clawhub)\r
- version\r
- notable risk signals — check for **any** of the following:\r
  - install count \x3C 10 (low adoption)\r
  - last update > 6 months ago (low maintenance signal)\r
  - requires broad file system or network permissions\r
  - published \x3C 2 weeks ago (very new, unvetted)\r
  - no author URL or verified identity\r
- replace/coexist plan\r
\r
Then explicitly ask for user confirmation.\r
\r
### 6) Install Execution\r
- `clawhub install \x3Cskill>`\r
- If replace approved: uninstall old first, then install new.\r
- If coexist: install and provide usage boundary guidance.\r
\r
### 7) Post-Install Report\r
Return:\r
- install/uninstall result\r
- final active skill set\r
- why this is best for the user goal\r
- follow-up checks\r
\r
## Output Format\r
\r
### Candidate Summary\r
- Name\r
- Purpose\r
- Source\r
- Version\r
- Install Count (if available)\r
- Link\r
\r
### Decision Summary\r
- Best choice (Top 1)\r
- Alternatives (Top 2/3)\r
- Replace/Coexist decision\r
- Why not other options\r
\r
### Pre-Install Confirmation\r
"Planned action: \x3Cinstall/replace/coexist>. Source: clawhub. Version: \x3C...>. Risks: \x3C...>. Proceed?"\r
\r
## Guardrails\r
\r
- Never install without explicit confirmation.\r
- Prefer fewer, higher-confidence recommendations over long noisy lists.\r
- If search quality is poor, refine keywords and re-run search before recommending.\r