Skip to main content

diagnostics

Use when troubleshooting build failures, test failures, daemon connectivity issues, or integration errors in Auto-Tundra.

Quellinformationen

Repository
ryanmaclean/tundra
Letzte Quellaktivität
22. Februar 2026 um 04:26
Erkannte Sprache von SKILL.md
Englisch
Sterne
0
Forks
0

Installationsoptionen

Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.

Quelldateien prüfen

Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.

SKILL.md wird angezeigt

SKILL.md
Quellanweisungen · Schreibgeschützte Vorschau
name
diagnostics
description
Use when troubleshooting build failures, test failures, daemon connectivity issues, or integration errors in Auto-Tundra.
allowed_tools
["Bash","Read"]
references
["/Users/studio/rust-harness/docs/PROJECT_HANDBOOK.md","/Users/studio/rust-harness/AGENTS.md"]
# Diagnostics ## Trigger Use this skill when something is broken: build errors, test failures, daemon won't start, API returns errors, or integrations fail. ## Diagnostic Ladder (follow in order) ### Step 1: Environment check ```bash # All-in-one health check at doctor -p /Users/studio/rust-harness -j # Manual checks rustc --version # Must be 1.91+ env | grep -E 'API_KEY|TOKEN' # Check credentials curl -sf http://localhost:9090/api/status | jq . # Daemon alive? ``` ### Step 2: Build diagnostics ```bash # Full workspace check (fast, no codegen) cargo check 2>&1 | head -50 # Single crate cargo check -p at-<crate> 2>&1 | head -50 # If linker errors on macOS: xcode-select --install ``` ### Step 3: Test diagnostics ```bash # Run with output visible cargo nextest run -p at-<crate> --no-capture 2>&1 | tail -80 # Run a single test cargo nextest run -p at-<crate> -E 'test(test_name)' --no-capture # Run with debug logging RUST_LOG=debug cargo nextest run -p at-<crate> --no-capture ``` ### Step 4: Daemon diagnostics ```bash # Start daemon with debug logging RUST_LOG=debug cargo run --bin at-daemon 2>&1 | tee /tmp/daemon.log # Check if ports are in use (API: 9090, UI: 3001) lsof -i :9090 -i :3001 # Kill stale daemon processes pkill -f at-daemon kill -9 $(lsof -t -i:9090 -i:3001) 2>/dev/null || true ``` ### Step 5: Integration diagnostics ```bash # GitLab curl -sf http://localhost:9090/api/gitlab/issues 2>&1 | jq . # Expected 503 with {"error":"...","env_var":"GITLAB_TOKEN"} if token missing # Linear curl -sf http://localhost:9090/api/linear/issues?team_id=X 2>&1 | jq . # GitHub curl -sf http://localhost:9090/api/github/issues 2>&1 | jq . ``` ### Step 6: Dependency/security diagnostics ```bash # Check for vulnerable dependencies cargo deny check 2>&1 | head -30 # Check for outdated dependencies cargo outdated -R 2>&1 | head -30 ``` ## Common Failure Patterns | Symptom | Likely Cause | Fix | |---------|-------------|-----| | `Connection refused :9090` | Daemon not running | `cargo run --bin at-daemon` | | `503 + env_var` in response | Missing credential | Ensure variables like `GITLAB_TOKEN` are `export`ed in the shell before starting the daemon, or modify `~/.config/auto-tundra/settings.toml`. | | `linking with cc failed` | Missing Xcode tools | `xcode-select --install` | | `cannot find crate` | Missing workspace member | Check `Cargo.toml` members | | Test timeout | Network-dependent test | Check `#[ignore]` gate or mock | | `port already in use` | Stale process | `lsof -i :9090 -i :3001` then `kill` | ## Rules 1. Always start at Step 1 and work down — don't jump to Step 5. 2. Capture output with `2>&1 | head -N` to avoid flooding context. 3. When reporting errors, include the exact error message and the command that produced it. 4. Never blindly `cargo clean` — it wastes 5-10 min of rebuild time. Only use if genuinely corrupt.
Auf GitHub ansehen