| name | claude-second-opinion |
| description | Non-Claude agents only. Get an independent second opinion from Claude CLI on decisions with meaningful trade-offs. |
Claude Second Opinion
If you are a Claude agent, STOP. Tell the user this skill is for non-Claude agents only.
Steps
- Build a concise prompt from
$ARGUMENTS if provided; otherwise infer the question from the current task context. Keep it focused on decision-relevant information:
You are providing a second opinion on a decision.
Context:
<context from the current task>
Question:
<the specific question or decision point>
Provide your independent analysis. If you disagree with the current approach, explain why and suggest alternatives.
- Run Claude CLI. Always set
--no-sandbox (nested sandbox-exec is prohibited on macOS):
If run_in_background and TaskOutput are available:
claude --no-sandbox -p "$(cat <<'CLAUDE_PROMPT'
<constructed prompt>
CLAUDE_PROMPT
)" 2>&1
Then retrieve with TaskOutput (block: true, timeout: 120000).
Otherwise (tmpfile + poll):
TMPFILE=$(mktemp /tmp/claude-opinion.XXXXXX.log)
claude --no-sandbox -p "$(cat <<'CLAUDE_PROMPT'
<constructed prompt>
CLAUDE_PROMPT
)" > "$TMPFILE" 2>&1 &
PID=$!
Poll $PID every 10-15s. Hard timeout: 300s. On timeout, kill $PID and clean up $TMPFILE.
-
Parse output: ignore startup logs, MCP errors, and reasoning traces before the final text. Treat the last plain-text block as the answer.
-
Present both perspectives (labeled by model), highlight agreements and disagreements. Evaluate Claude's suggestion critically — do not blindly adopt it.
Error Handling
On any failure — claude not found, auth error, network error, or timeout — report the reason to the user and proceed with your own analysis only. Discard partial output on timeout as unreliable.
Shell Escaping
Use a heredoc (<<'CLAUDE_PROMPT') to build prompt text safely. This avoids shell escaping and quoting issues for multi-line content. Note that prompts passed via -p are still subject to ARG_MAX.
