Research-aligned self-consistency for debugging. Spawns independent solver agents that each explore and debug the problem from scratch. Uses majority voting. Based on "Self-Consistency Improves Chain of Thought Reasoning" (Wang et al., 2022).
SkillScore80Maintained
Automatic
◆Coding pack
Last verified 8/3/2026
What it changes
Passed 16 of 20 test cases with the skill, and 7 of 20 without it.
Without skill36%
With skill80%
Checks it passed that the model alone missed
States that the allowed agent count range is between 3 and 10
Indicates that the agent count choice must be between 3 and 10
Lists the isolated exploration tools as Read, Grep, Glob, and LS
and 13 more in the full report
14 fixed6 passed either way2 still failing3 broke
Pass-rate lift
+44 pp
Tokens
+81%
Agent turns
0%
Response time
-40%
Model gemini-3.6-flashMeasured 8/3/2026v120 of 25 cases passed with skill · 20 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 11 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 11
Prompt
In a debug council run with 7 agents for `Database lock timeout on user deletion`, 5 agents proposed adding a transactional retry decorator (Approach A), while 2 agents proposed increasing pool size (Approach B). What confidence level should be reported for Approach A?
What was checked, without and with the skill
Checked
Without skill
With skill
Rates the confidence level as HIGH because Approach A secured 50% or more of the total votes
Missed
Passed
case-11 · gemini-3.6-flash · 8/3/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md262 linessingle file
Debug Council: Research-Aligned Self-Consistency
Pure implementation of self-consistency (Wang et al., 2022). Each agent receives the raw user prompt and explores/debugs independently. No pre-processing, no shared context. Majority voting selects the answer.
Use this for bugs and problems with ONE correct answer.
Step 0: Ask User How Many Agents
Before doing anything else, ask the user how many solver agents to use:
How many debug agents would you like me to use? (3-10)
Recommendations:
- 3 agents: Faster, still reliable
- 5 agents: Good balance
- 7 agents: High confidence
- 10 agents: Maximum confidence (critical bugs)
Note: Each agent will independently explore the codebase and find the bug.
This takes longer but provides true independence per the research.
Wait for the user's response. If they specified a number (e.g., "debug council of 5"), use that.
Minimum: 3 agents | Maximum: 10 agents
CRITICAL: Pure Research Alignment
What This Means
NO orchestrator exploration - Do NOT read files or gather context before spawning agents
Raw user prompt to all agents - Each agent gets the user's original request, unchanged
Each agent explores independently - Agents discover the codebase themselves
True independence - No shared context, no cross-contamination