v0.3.1

Metadata-only release. Republishes the package so the README on PyPI picks up the corrected install instructions (uv tool install claude-atlas) and the absolute logo URL. No code changes vs 0.3.0.

v0.3.0 — 'check' command for CI

New: claude-atlas check

A lint-style health check command designed for pre-commit hooks, CI pipelines, and quick terminal use. No HTML rendering — just fast, scriptable output.

Output formats

• text (default): grouped by source file, with severity, detail, and 💡 suggested fix
• json: machine-readable summary + issues array
• github: ::error / ::warning / ::notice annotations for GitHub Actions, with proper escaping

Exit codes

• 0: no issues at-or-above --max-severity (default: high)
• 1: failing issues found
• 2: invalid arguments or scan error

Flags

• --max-severity {low,medium,high,none} — threshold for exit 1. none means report-only.
• --format {text,json,github}
• --top N — show top N most severe (0 = all). Default 10.
• --quiet/-q — only print summary line.

Examples

# Default: fail on HIGH-severity issues
claude-atlas check

# Pre-commit, including MEDIUM
claude-atlas check --max-severity medium --quiet

# GitHub Actions
claude-atlas check --format github

# Pipe to jq
claude-atlas check --format json --top 0 | jq '.summary'

Full changelog: v0.2.0...v0.3.0

v0.2.0 — Signal + UX overhaul

Signal

• Severity classification (high/medium/low) on every issue, surfaced in the sidebar, graph edges, and stats.
• Domain stopwords drop noise from trigger collisions. Generic tokens like agent, user, task, proactively no longer cause false positives — this was the main noise source in v0.1.x.
• Short tokens (< 5 chars) excluded from collision detection. A 72-agent real-world scan went from ~100 issues to ~17 meaningful ones.

UX

• Search box filters nodes, issues, and orphans live.
• Severity filters alongside kind filters.
• Grouped issues by severity → kind with accordion; HIGH expanded by default.
• Bidirectional click: clicking an issue in the sidebar highlights the pair on the graph and zoom-fits them. Clicking a node still populates the detail pane.
• Hover tooltips on nodes show a quick description preview.
• Isolated artifacts get their own tab and a dashed border on the graph — for CLAUDE.md files in repos without a .claude/, or skills with no overlap.
• Suggested fixes (💡 text hints) on every issue. No automation — you still decide.

Breaking

• Edge dataclass gains a severity field. External consumers of ScanResult.to_dict() will see a new severity key.

Full changelog: v0.1.1...v0.2.0

v0.1.1 — Frontmatter fallback + readable layout

Fixed

• Multi-line description: frontmatter values (common in real-world agents like studio-coach) were silently parsed as empty metadata, making those artifacts invisible to trigger heuristics. Parser now falls back to regex extraction when YAML fails.

Improved

• Graph layout: node repulsion, edge length, and component spacing tuned for 70+ artifact scans.
• Labels: ellipsis wrap at 120px (240px when selected), so names no longer overlap.
• Issue edges rendered with higher opacity and thicker stroke.

Notes

Expect more trigger_collision issues in the Issues tab than v0.1.0. This is correct — previously-invisible agents are now being analyzed.

Full changelog: v0.1.0...v0.1.1