Skip to content

Wolfe-Jam/faf-mcp

v2.1.0 Breaking

This release includes 2 breaking changes for platform teams planning a safe upgrade.

Published 16d MCP Developer Tools
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agents-md ai ai-context claude cline cursor
+14 more
developer-tools faf gemini grok mcp mcp-server model-context-protocol nodejs npm project-dna typescript vscode warp windsurf

Summary

AI summary

All commands now single-source scoring and rendering from faf-cli, eliminating fabricated scores.

Changes in this release

Feature Medium

faf_display now single-sources faf-cli's generateProjectHtml for consistent project.html output.

faf_display now single-sources faf-cli's generateProjectHtml for consistent project.html output.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: low

Dependency Medium

faf-cli ^6.7.1 now consumes faf-cli's typed public API as the single source of truth.

faf-cli ^6.7.1 now consumes faf-cli's typed public API as the single source of truth.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: high

Bugfix Medium

Removed fabricated file-presence pseudo-score, non-deterministic timestamp, off-canon colors, and divergent template in faf_display.

Removed fabricated file-presence pseudo-score, non-deterministic timestamp, off-canon colors, and divergent template in faf_display.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: low

Bugfix Medium

faf_show, faf_score, and faf_status now route through getFafScore() using faf-cli's scorer and canonical tier ladder.

faf_show, faf_score, and faf_status now route through getFafScore() using faf-cli's scorer and canonical tier ladder.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: low

Refactor Medium

Removed fabricated full scorecard and its unmeasured performance claims from faf_show, faf_score, faf_status.

Removed fabricated full scorecard and its unmeasured performance claims from faf_show, faf_score, faf_status.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: low

Full changelog

Truthful, single-sourced scoring. faf-mcp no longer computes its own
score — every score surface reads faf-cli's real scorer, the same
number your AI and faf score see.

Fixed

  • faf_display — rendered a fabricated file-presence pseudo-score
    (40/30/15/14, max 99), a non-deterministic timestamp, off-canon
    colors and a divergent template. Now single-sources faf-cli's
    generateProjectHtml — byte-identical to faf show / faf export --html. Default output project.html (gitignored — a view, not
    source).
  • faf_show / faf_score / faf_status — same fake file-presence
    score, plus the banned medal / colored-circle tier ladder
    (🥇🥈🥉🟢🟡🔴🤍). All routed through one getFafScore() helper
    backed by faf-cli's scorer + canonical tier ladder. Deterministic
    output (no timestamps, no rotating quotes). Removed the fabricated
    full scorecard and its unmeasured performance claims.

Changed

  • faf-cli ^6.7.1 — consumes faf-cli's typed public API (the real
    scorer + the project.html renderer) as the single source of truth.

299 tests passing (8 suites). Single-sourced from faf-cli ^6.7.1.

FAF defines. MD instructs. AI codes.

Breaking Changes

  • Removed fabricated file‑presence pseudo‑score and related non‑deterministic behavior from `faf_display`, `faf_show`, `faf_score`, and `faf_status`.
  • Changed minimum faf-cli version to ^6.7.1; all commands now use its typed public API for scoring and project.html rendering.

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

Track Wolfe-Jam/faf-mcp

Get notified when new releases ship.

Sign up free

About Wolfe-Jam/faf-mcp

Universal persistent project context for Cursor, Windsurf, Cline, VS Code, and all MCP-compatible platforms (including Claude Desktop). IANA-registered format (application/vnd.faf+yaml). 17 native tools, AI-readiness scoring.

All releases →

Beta — feedback welcome: [email protected]