This release adds 3 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Summary
AI summaryAdded comprehensive MCP server instructions and two new tools: evidence scoring and decision templates.
Full changelog
What's New in v0.2.6
MCP Server Instructions
Added comprehensive server-level instructions following the Model Context Protocol specification. These instructions help AI clients understand how to optimally use the Counsel tools:
- Critical Workflow: Documents the recommended flow (sharpen → start → poll → get report)
- Mode Selection: Explains quick, standard, deep, and research modes
- Phase Constraints: Details debate phase progression (diverge → attack → crux → integrate)
- Support Tools: Lists available support tools and their purposes
This enhancement improves the developer experience by providing clear guidance directly to AI clients through the MCP protocol.
New Tools
Evidence Scoring (score_evidence_pack):
- Score evidence coverage and credibility before running analysis
- Returns breakdown by coverage, credibility, and relevance
- Identifies gaps in evidence
- Provides recommendations for improvement
Decision Templates:
list_decision_templates: List available decision templates for your workspaceuse_decision_template: Instantiate templates with user-provided variables- Supports both workspace-specific and public templates
- Returns fully instantiated questions ready for consultation
Improvements
- Fixed version consistency across all files (package.json, src/cli.ts, src/index.ts)
- Added CHANGELOG.md for tracking version history
Installation:
npm install [email protected]
Full Changelog: https://github.com/mercurialsolo/counsel-mcp/compare/v0.2.5...v0.2.6
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About mercurialsolo/counsel-mcp
Connect AI agents to the Counsel API for strategic reasoning, multi-perspective debate analysis, and interactive advisory sessions.
Related context
Beta — feedback welcome: [email protected]