[lxsaah]

Description

aimdb used to exit with 1 for every error, so a script couldn't tell an unreachable instance (worth retrying) from a refused request. Every command now uses one exit-code scheme:

Code	Class
0	Success
1	Rejected: the instance or a remote service refused the request (RecordNotFound, WriteProtected, ServerError, Other)
2	Usage: clap's usage errors, UnsupportedEndpoint, InvalidJson
3	Environment: ConnectionFailed, NoInstancesFound, ProtocolMismatch, Io, Json
4+	Command-specific (e.g. join's device flow, design 056 §5)

Changes

• CliError::exit_code() with the constants EXIT_REJECTED, EXIT_USAGE and EXIT_ENVIRONMENT. The match has no wildcard, so a new variant won't compile until its code is chosen.
• main exits with e.exit_code().
• Design 056 §5 now refers to this shared scheme instead of "1 for all existing variants", and step 1 of §9 is marked done.
• The CLI README gets an "Exit codes" section, and the CLI changelog a Changed (breaking) entry: connection errors now exit with 3 instead of 1.

Tests

• Unit test: every CliError variant maps to its class.
• tests/exit_codes.rs runs the aimdb binary: an unreachable unix:// socket → 3, --connect bogus://x → 2, invalid JSON for record set --dry-run → 2, an unknown subcommand → 2. The first three fail when main still exits with exit(1).

Verified locally: cargo test -p aimdb-cli and cargo clippy -p aimdb-cli --all-targets. The full make check is left to CI.

Found while doing this, but not changed here: aimdb-cli doesn't compile with --features yaml. That gets its own PR.

Related Issue

• Closes [FEAT] aimdb CLI: distinct exit codes per failure class #334

Checklist

• I have read the CONTRIBUTING.md document.
• My code follows the project's coding standards.
• I have added tests to cover my changes.
• All new and existing tests passed (make check). Left to CI.
• I have updated the documentation accordingly.

🤖 Generated with Claude Code

https://claude.ai/code/session_01YPDjQ947yKjLLAq3Xzgu3e