Smoke Test Checklist

Quick pass to verify everything works after a build. Run these in order – takes ~15 minutes.

Phase 1: Build & Boot (no Ollama needed)

#	Command	Pass if
1	`cargo build`	Compiles without errors
2	`cargo test`	~645 tests pass (workspace sum; see `docs/status/current-state.md`)
3	`cargo run -- --help`	Shows help with all subcommands
4	`cargo run -- session create smoke-test`	Prints session UUID
5	`cargo run -- session list`	Shows the smoke-test session
6	`cargo run -- finance status`	Prints status (no crash)
7	`cargo run -- code analyze .`	Prints analysis results
8	`cargo run -- harvest pipeline`	Prints pipeline (all zeros OK)

Phase 2: Web Server (no Ollama needed)

Start cargo run in a terminal, then in another:

#	Command	Pass if
9	`curl -s localhost:3000/api/health`	Returns 200
10	`curl -s localhost:3000/api/departments \| jq length`	Returns 14
11	`curl -s localhost:3000/api/agents \| jq length`	Returns >= 7 (seeded)
12	`curl -s localhost:3000/api/skills \| jq length`	Returns >= 5 (seeded)
13	`curl -s localhost:3000/api/rules \| jq length`	Returns >= 5 (seeded)
14	`curl -s localhost:3000/api/config/tools \| jq length`	Returns >= 22
15	`curl -s localhost:3000/api/db/tables \| jq length`	Returns >= 5
16	Open `http://localhost:3000` in browser	Dashboard renders

Phase 3: LLM Features (requires Ollama)

#	Command	Pass if
17	`cargo run -- forge mission today`	Streams daily plan
18	`curl -N localhost:3000/api/chat -H 'Content-Type: application/json' -d '{"message":"hi"}'`	SSE stream with text deltas
19	`curl -N localhost:3000/api/dept/code/chat -H 'Content-Type: application/json' -d '{"message":"list files"}'`	SSE stream with tool calls
20	`cargo run -- content draft "test"`	Streams content draft

Phase 4: Interactive Surfaces

#	Action	Pass if
21	`cargo run -- shell` -> type `help` -> type `exit`	REPL starts, shows help, exits cleanly
22	`cargo run -- shell` -> `use finance` -> `status` -> `back`	Context switching works
23	`cargo run -- --tui` -> press `q`	TUI renders 4 panels, exits cleanly

Phase 5: CRUD Cycle

Pick any entity (agents, skills, rules, hooks, workflows) and verify the full lifecycle:

#	Action	Pass if
24	`POST` create	Returns 200 with ID
25	`GET` by ID	Returns created entity
26	`GET` list	Entity appears in list
27	`PUT` update	Returns updated entity
28	`GET` by ID again	Shows updated fields
29	`DELETE`	Returns 200/204
30	`GET` by ID again	Returns 404

Example CRUD cycle (agents):

# Create
ID=$(curl -s -X POST localhost:3000/api/agents \
  -H 'Content-Type: application/json' \
  -d '{"name":"smoke-agent","role":"Test","instructions":"test"}' | jq -r '.id')
echo "Created: $ID"

# Read
curl -s localhost:3000/api/agents/$ID | jq .name
# Expected: "smoke-agent"

# Update
curl -s -X PUT localhost:3000/api/agents/$ID \
  -H 'Content-Type: application/json' \
  -d '{"name":"updated-agent","instructions":"updated"}' | jq .name
# Expected: "updated-agent"

# List (verify it's there)
curl -s localhost:3000/api/agents | jq '.[].name' | grep updated-agent
# Expected: "updated-agent"

# Delete
curl -s -X DELETE localhost:3000/api/agents/$ID -w "\n%{http_code}\n"
# Expected: 200 or 204

# Verify gone
curl -s -o /dev/null -w "%{http_code}" localhost:3000/api/agents/$ID
# Expected: 404

Automated Test Suite

For reference, the full automated suite:

# Full workspace
cargo test

# Individual crates
cargo test -p rusvel-core
cargo test -p rusvel-db
cargo test -p rusvel-api
cargo test -p rusvel-llm
cargo test -p forge-engine
cargo test -p content-engine
cargo test -p harvest-engine
cargo test -p code-engine
cargo test -p flow-engine
cargo test -p rusvel-tool
cargo test -p rusvel-agent

# Line coverage
./scripts/coverage.sh

# Benchmark
cargo bench -p rusvel-app --bench boot

Keyboard shortcuts

RUSVEL