fix(mcp): parse --json=false and --exit-code=false as real booleans

[rsnetworkinginc]

Closes #8689

Problem

parseOptions' generic --key=value inline-equals handler (added for --format=table, #2231) stores the raw string for every flag. So --json=false set options.json = "false" — a non-empty, truthy string — and every options.json consumer (23 call sites) plus cli-error.ts's argsWantJson checks truthiness, so the flag turned JSON output ON, the opposite of what it says. The identical bug hit doctor --exit-code=false, which still escalated the exit code to 1.

Fix

The Requirements allow "the same === true (or equivalent explicit boolean-parsing) pattern" — this PR implements the explicit boolean parse at the single choke point every consumer flows through, so all 23 options.json consumers and doctor's options.exitCode check compare against a guaranteed real boolean (never a raw string), exactly the semantics of the --refresh/--help === true convention:

• parseOptions (packages/loopover-mcp/bin/loopover-mcp.ts): the inline --key=value values for the boolean flags json/exitCode now parse to a real boolean — only the exact string "false" (and empty "", which was already falsy) disables; any other value keeps today's enabling behavior. Bare --json and --json=true are unchanged; only the =false form changes behavior (now correctly disabling).
• doctor: the exit-code check is now the explicit options.exitCode === true && … comparison (the --refresh line-4033 convention), so --exit-code=false keeps the exit code at 0 even when checks need attention.
• argsWantJson (packages/loopover-mcp/lib/cli-error.ts): --json=false no longer counts as wanting JSON on the failure path.

Tests (in-process runCli, the mcp-cli-basics.test.ts #8587 pattern)

New test/unit/mcp-cli-bool-flag-parsing.test.ts + an argsWantJson case in test/unit/mcp-cli-error.test.ts, all red on main before the fix:

Deliverable	Test	Before (main)	After
whoami --json=false → plain text	asserts output is exactly JSONbored\n, not JSON	failed: emitted {"status":"authenticated","l…	passes
doctor --exit-code=false vs failing auth check → exit 0	asserts in-process return 0 with status: "needs_attention"	failed: expected 1 to be +0	passes
bare --json / --json=true still enable JSON	parses whoami JSON payload for both forms	passed	passes (regression guard)
bare --exit-code / --exit-code=true still escalate	asserts return 1 vs failing check	passed	passes (regression guard)
argsWantJson("--json=false") → false	direct unit test	failed: expected true to be false	passes
non-boolean inline values stay raw strings	init-client --print=codex renders the codex snippet	passed	passes

Full runs: mcp-cli-bool-flag-parsing, mcp-cli-error, mcp-cli-basics, mcp-cli-doctor — 51/51 passing; tsc --noEmit clean.

Coverage

Every changed executable line in bin/loopover-mcp.ts (the booleanFlags set, the inline boolean parse, doctor's return) and lib/cli-error.ts is exercised in-process with both branch directions taken (verified against the lcov report: all BRDA arms on the changed lines are non-zero), covering the =false and default/=true branches for both flags as required.

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.82%. Comparing base (abf88ab) to head (c671397).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8725      +/-   ##
==========================================
- Coverage   90.56%   83.82%   -6.74%     
==========================================
  Files          96       98       +2     
  Lines       22490    24757    +2267     
  Branches     3884     4744     +860     
==========================================
+ Hits        20367    20752     +385     
- Misses       1945     3727    +1782     
- Partials      178      278     +100     

Flag	Coverage Δ
backend	16.98% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
packages/loopover-mcp/bin/loopover-mcp.ts	16.72% <100.00%> (ø)
packages/loopover-mcp/lib/cli-error.ts	100.00% <100.00%> (ø)

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-26 01:15:26 UTC}

4 files · 1 AI reviewer · no blockers · readiness 98/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This PR fixes a real bug: the generic inline `--key=value` handler in `parseOptions` stored raw strings, so `--json=false` and `--exit-code=false` became truthy strings and silently enabled behavior opposite of what was requested. The fix scopes an explicit boolean parse to the `json`/`exitCode` flags at the single choke point, updates `doctor`'s exit-code check to an explicit `=== true` comparison consistent with the existing `--refresh` convention, and patches `argsWantJson` to stop treating `--json=false` as wanting JSON. The diff is narrow, well-targeted, and accompanied by tests that exercise the real CLI dispatcher in-process (not fabricated payloads), covering both the previously-broken `=false` cases and the unchanged bare/`=true` cases.

Nits — 5 non-blocking

• packages/loopover-mcp/bin/loopover-mcp.ts: the new `booleanFlags` Set uses camelCase keys ("json", "exitCode") consistent with `camel(inlineKey)` above it, but double-check that `--exit-code=false` actually maps through `camel()` to `"exitCode"` before reaching the `booleanFlags.has` check — worth a one-line comment confirming this since it's easy to typo the camelCase key and silently miss the flag.
• test/unit/mcp-cli-bool-flag-parsing.test.ts: this new file duplicates a fair amount of harness boilerplate (`withEnv`, `captureStdout`, fixture server setup) that likely already exists in `mcp-cli-basics.test.ts`/`mcp-cli-doctor.test.ts` per the PR description — consider extracting shared helpers if this pattern repeats a third time.
• The external brief's 'long file' flag on `loopover-mcp.ts` (5955 lines) is pre-existing file size, not something this diff introduces — not actionable for this PR.
• packages/loopover-mcp/bin/loopover-mcp.ts:5495: consider a short inline comment or test asserting `camel("exit-code") === "exitCode"` matches the `booleanFlags` Set entry, since a future rename of either could silently break the disable path with no type error to catch it.
• The fix correctly scopes to `json`/`exitCode` only, per the `--print=codex` passthrough test in mcp-cli-bool-flag-parsing.test.ts:127-131 — this guards against over-broad boolean coercion of unrelated inline flags, good call.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#8689
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (1 linked issue).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 55 registered-repo PR(s), 22 merged, 3 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	rsnetworkinginc; Gittensor profile; 55 PR(s), 3 issue(s).
Improvement	✅ Minor	risk: clean · value: minor · LLM: moderate

Linked issue satisfaction

Addressed
The diff fixes the inline `--key=value` parser to coerce json/exitCode to real booleans (only "false"/empty disable), updates doctor's exit-code check to `=== true`, and fixes argsWantJson, matching the requested convention; it also adds tests covering --json=false, --json/--json=true, --exit-code=false, and --exit-code/--exit-code=true.

Review context

• Author: rsnetworkinginc
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: TypeScript
• Official Gittensor activity: 55 PR(s), 3 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Keep the PR focused and include validation evidence before maintainer review.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by LoopOver, a quiet PR intelligence layer for OSS maintainers.

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.