feat: add structured exit codes for CI/CD scripting

[mostafa]

Summary

• Introduce categorized exit codes (0 SUCCESS, 1 FINDINGS, 2 RULE_ERROR, 3 CONFIG_ERROR) so CI/CD pipelines can distinguish "the tool found something" from "the tool broke."
• Add eval --fail-on-detection flag to exit 1 when any detection or correlation fires.
• Add lint --fail-level <error|warning|info> flag to control the severity threshold that triggers a non-zero exit.
• Replace all blanket process::exit(1) calls across every subcommand (eval, lint, validate, convert, parse) and the daemon with the appropriate categorized code.

Test results

All tests verified manually on feat/structured-exit-codes @ 9453f86.

Automated

• cargo clippy --workspace --all-targets --all-features -- -D warnings -- clean
• cargo test --workspace -- all passing

Manual: eval --fail-on-detection

Test	Command	Expected	Actual
Match + flag	rsigma eval --fail-on-detection -r rule.yml -e '{"CommandLine":"mimikatz.exe …"}'	exit 1	exit 1
No match + flag	rsigma eval --fail-on-detection -r rule.yml -e '{"CommandLine":"whoami"}'	exit 0	exit 0
Match, no flag	rsigma eval -r rule.yml -e '{"CommandLine":"mimikatz.exe …"}'	exit 0	exit 0
Match + flag + @file	rsigma eval --fail-on-detection -r rule.yml -e @match.json	exit 1	exit 1

Manual: lint --fail-level

Test	Command	Expected	Actual
Default (error)	rsigma lint rule.yml (has warnings + infos, no errors)	exit 0	exit 0
--fail-level warning	rsigma lint --fail-level warning rule.yml	exit 1	exit 1
--fail-level info	rsigma lint --fail-level info rule.yml	exit 1	exit 1

Manual: categorized error codes

Test	Command	Expected	Actual
Bad rule syntax	rsigma eval -r bad_syntax.yml -e '{…}'	exit 2 (RULE_ERROR)	exit 2
Nonexistent pipeline	rsigma eval -r rule.yml -p nonexistent.yml -e '{…}'	exit 3 (CONFIG_ERROR)	exit 3