Local Council Execution

Fallback council of blind Claude subagents when no external AI providers are configured

Works with setup

Test report

Verdict
Works with setup
Score
8.0/10
Tested
Jul 21, 2026
Environment
Claude Code 2.x (agent harness)
Upstream re-checked
Aug 10, 2026 · fa9cfdc

Cloned the repo to a temp dir, set CLAUDE_PLUGIN_ROOT, and exercised the machinery under bash. query-council.sh --list-default returned empty (no providers = the legitimate local-council condition). Sourcing lib/roles.sh + lib/prompts.sh, I confirmed local_council_roles "" gives devil,simplicity,security,scalability, count 5/3 extend/trim correctly, count 99 clamps to all 8 roles, get_role_name devil='Devil's Advocate', validate_roles rejects unknown roles, and build_prompt_with_role injected the role prompt + user question exactly as the template documents. All 3 spot-checked referenced files returned HTTP 200. No security smells (no curl|sh, base64, or exfiltration). Could NOT measure output: Step 3 (spawn N parallel, blind, background general-purpose subagents) is the load-bearing mechanism and I have no subagent-spawn tool from this context, so no council artifact was produced.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

  • Installs cleanly 5/5
  • Triggers reliably 5/5
  • Output vs. baseline 5/10
  • Docs & honesty 5/5

What Local Council Execution does

Runs a local fallback council when no external AI providers are configured, spawning several independent Claude subagents that each answer from one assigned role, then synthesizing the spread of perspectives. It repeatedly stresses that same-model agreement is weak signal, not cross-vendor corroboration. Triggers via the ask command's --local flag or after the user accepts the local-council offer when no provider keys or CLIs are found.

How to install Local Council Execution

git clone --depth 1 https://github.com/hex/claude-council.git /tmp/local-council-execution-src
mkdir -p ~/.claude/skills
cp -R /tmp/local-council-execution-src/skills/local-council-execution ~/.claude/skills/local-council-execution
# CAVEAT: this skill calls plugin scripts via ${CLAUDE_PLUGIN_ROOT} (config/roles.json, scripts/lib/prompts.sh,
# scripts/lib/roles.sh, scripts/query-council.sh). Those resolve ONLY when claude-council is installed as a PLUGIN.
# A bare skill copy has SKILL.md but not the scripts. For a working setup install the whole plugin instead:
#   /plugin marketplace add hex/claude-council   then   /plugin install claude-council
# Requires: jq. Also requires a subagent-spawn (Agent/Task) tool at runtime to launch the parallel members.
# Local council is a fallback invoked by `/claude-council:ask --local` when no providers are configured.

Skills live in ~/.claude/skills/ (global) or .claude/skills/ (per-project). Restart Claude Code after installing.

Commands — how to trigger Local Council Execution

  • /local-council-execution Fallback council of blind Claude subagents when no external AI providers are configured

It also activates on plain-language prompts like these:

  • Run a local council review on this decision
  • Get five expert perspectives on this plan
  • Ask the local council since no providers are set up

Frequently asked questions

Is the Local Council Execution skill free?
Yes. The skill itself is free from hex/claude-council. SkillProof publishes the install command and an independent test verdict at no cost.
Does Local Council Execution work with Claude Code?
We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Works with setup. Cloned the repo to a temp dir, set CLAUDE_PLUGIN_ROOT, and exercised the machinery under bash. query-council.sh --list-default returned empty (no providers = the legitimate local-council condition). Sourcing lib/roles.sh + lib/prompts.sh, I confirmed local_council_roles "" gives devil,simplicity,security,scalability, count 5/3 extend/trim correctly, count 99 clamps to all 8 roles, get_role_name devil='Devil's Advocate', validate_roles rejects unknown roles, and build_prompt_with_role injected the role prompt + user question exactly as the template documents. All 3 spot-checked referenced files returned HTTP 200. No security smells (no curl|sh, base64, or exfiltration). Could NOT measure output: Step 3 (spawn N parallel, blind, background general-purpose subagents) is the load-bearing mechanism and I have no subagent-spawn tool from this context, so no council artifact was produced.
What is the Local Council Execution SkillProof Score?
8.0/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 5/10, docs & honesty 5/5.
How do I install Local Council Execution?
Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use Local Council Execution with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.