Skip to main content

diagnostics

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

Informations de source

Dépôt
ryanmaclean/tundra
Dernière activité de la source
22 février 2026 à 04:26
Langue détectée de SKILL.md
anglais
Étoiles
0
Forks
0

Options d'installation

Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.

Vérifiez les fichiers source

Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.

Affichage de SKILL.md

SKILL.md
Instructions source · Aperçu en lecture seule
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.
Voir sur GitHub