docs(roadmap): add #464 — --output-format value parsing strict-match-only with 5 compounding gaps (case-hostile, whitespace-hostile, no Did-you-mean, no empty branch, classifier orphan); second confirmed instance of #463 classifier-orphan pattern

ROADMAP.md

@@ -6422,3 +6422,62 @@ Original filing (2026-04-18): the session emitted `SessionStart hook (completed)
 
 448. **`sandbox --output-format json` has contradictory state flags — `enabled:true, supported:false, active:false, filesystem_active:true, allowed_mounts:[]`: claim that sandbox is "enabled" while OS doesn't support namespace isolation and `allowed_mounts:[]` is empty contradicts `filesystem_active:true filesystem_mode:"workspace-only"`** — dogfooded 2026-05-11 by Jobdori on `7244a82b` in response to Clawhip pinpoint nudge at `1503403842920779917` (using fresh-current-main runner at `/tmp/claw-dog-1430` per gajae's 14:00 protocol switch). Reproduction: `claw sandbox --output-format json` on macOS (where `unshare` is unavailable) returns `{"active":false,"active_namespace":false,"active_network":false,"allowed_mounts":[],"enabled":true,"fallback_reason":"namespace isolation unavailable (requires Linux with \`unshare\`)","filesystem_active":true,"filesystem_mode":"workspace-only","in_container":false,"kind":"sandbox","markers":[],"requested_namespace":true,"requested_network":false,"supported":false}`. **Three contradictions in the same envelope:** (a) `enabled:true` AND `supported:false`: what does "enabled" mean if the OS doesn't support sandboxing? Read literally, sandbox is *enabled but unsupported* — semantic nonsense. The likely intent is "user requested sandbox in config" but the field name `enabled` says "is ON". A better name would be `requested:true` or `config_intent:true`, with `enabled` reserved for the actually-active state. (b) `filesystem_active:true, filesystem_mode:"workspace-only"` AND `allowed_mounts:[]`: if the filesystem fence is active in workspace-only mode, the workspace directory itself MUST be an allowed mount. An empty `allowed_mounts:[]` array combined with `filesystem_active:true` means either (i) the fence is being misreported (it's not really active), (ii) the workspace is implicit and `allowed_mounts` only lists *additional* mounts, or (iii) the fence has no allowed paths and nothing is readable — all three are inconsistent with the user-facing summary. (c) `active:false` AND `filesystem_active:true`: the top-level `active` field is a single boolean summary, but it disagrees with `filesystem_active:true` (one component is active). Either `active` is "all components active" (then it should be `false` when any component is off) or "any component active" (then it should be `true` when filesystem is). The current value is `false` despite filesystem being active. **Sibling: no `claw sandbox --help`**: `claw sandbox status` and `claw sandbox --help` go to LLM-prompt fallback or hang (gajae confirmed at 13:00 that `sandbox status` returns typed `cli_parse` but `sandbox --help` is bounded — schema is non-uniform across help paths). **Required fix shape:** (a) rename `enabled` to `requested` or `config_intent` to disambiguate from "currently active"; (b) make `allowed_mounts` explicitly include the workspace when filesystem_mode is "workspace-only" (`allowed_mounts:[{path:"<cwd>",writable:true,reason:"workspace_root"}]`); (c) document the `active` aggregate semantics: pick either "all" or "any" composition rule and document the choice; (d) add `active_components:["filesystem"]` array as a richer alternative to the single boolean — surfaces exactly which sandbox subsystems are live; (e) regression test: when `filesystem_mode == "workspace-only"`, `allowed_mounts` MUST contain the cwd and `active` must agree with the documented composition rule. **Why this matters:** sandbox is the trust surface — automation that checks `sandbox.active == true` before running a risky LLM prompt sees `false` (no namespace, no network) and assumes no isolation, but `filesystem_active:true` means there IS partial isolation. The mixed signal forces consumers to OR all `*_active` fields together. Cross-references #428 (default permission_mode=danger-full-access — paired with sandbox-not-active means zero isolation), #444 (no broad-cwd guard — sandbox is the only safety net and its status is unclear). Source: Jobdori live dogfood, `7244a82b`, 2026-05-11.
 
+
+449. **`claw session list --output-format json` routes through `CliAction::ResumeSession` and hits the auth gate, returning `kind:"missing_credentials"` — but `session list` is a pure local filesystem read that requires no API credentials; by contrast, `claw session` (without `list`) correctly short-circuits with `kind:"unknown"` + "is a slash command" message without touching the auth gate** — dogfooded 2026-05-12 by Jobdori on `8f55870d` in response to Clawhip pinpoint nudge at `1503638404842131456`. Reproduction (no creds, isolated env): `env -i HOME=$HOME PATH=$PATH claw session list --output-format json` → `{"error":"missing Anthropic credentials...","kind":"missing_credentials"}` exit 1. `env -i HOME=$HOME PATH=$PATH claw session --output-format json` → `{"error":"`claw session` is a slash command...","kind":"unknown"}` exit 1 (no auth check). Root cause: the parser routes `session list` via `parse_resume_session_args` treating `list` as a session-path token, producing `CliAction::ResumeSession { session_path: "list", commands: [] }`. `resume_session()` then calls `LiveCli::new()` which instantiates the Anthropic client and fires the credentials guard. The `SlashCommand::Session { action: Some("list") }` special-case path in `run_resume_command()` (line 3654 comment: "`/session list` can be served from the sessions directory without a live session") is only reachable after auth passes — the no-creds guard fires before the slash-command dispatch loop. **Asymmetry:** the internal code already knows `session list` is credential-free (the comment at line 3654 says so), but the CLI entrypoint forces creds before the command ever reaches that branch. **Sibling: `session list` with no sessions returns `kind:"session_load_failed"` (from `--resume latest` fallback) rather than `{"kind":"session_list","sessions":[],"session_details":[]}` — the empty-sessions case is misrouted to the resume-failure path instead of a list-success with zero entries.** **Required fix shape:** (a) add a dedicated `CliAction::SessionList { output_format }` variant dispatched when `claw session list` is parsed — do not route through `ResumeSession`; (b) implement `run_session_list(output_format)` as a credentials-free function that calls `list_managed_sessions()` directly (same logic as the slash-command special-case at line 3659); (c) ensure empty sessions returns `{"kind":"session_list","sessions":[],"session_details":[],"active":null}` with exit 0, not a `session_load_failed` error; (d) add the same fix for sibling local-only commands that currently hit the auth gate: `session delete <id>`, `session export <id>`; (e) regression test: `claw session list --output-format json` with no credentials returns `kind:"session_list"` exit 0. **Why this matters:** session list is the canonical inventory surface for automation pipelines — `claw session list --output-format json | jq '.session_details[] | .id'` is the idiomatic way to enumerate sessions for replay, export, or resume. Requiring API credentials to read a local directory listing breaks offline use, CI environments with no API key configured, and any scripting that runs before credential setup. Cross-references #357 (session list requires creds — this is the same bug surfaced by that entry; #449 provides the root-cause path trace), #369 (session help/fork require creds), #427 (resume --help hits auth gate), #431 (skills uninstall requires creds). Source: Jobdori live dogfood, `8f55870d`, 2026-05-12.
+
+
+
+450. **`prompt` emits `kind:"missing_credentials"` JSON on STDERR (not stdout), leaving stdout at 0 bytes — automation pattern `output=$(claw prompt hello --output-format json)` captures nothing on auth-absent failure; `doctor` correctly surfaces `auth.status:"warn"` with `api_key_present:false` but exposes no `prompt_ready:false` field that automation can check before invoking `prompt`** — dogfooded 2026-05-16 by Jobdori on `a35ee9a0` in response to Clawhip pinpoint nudge at `1505208225321062521`. Exact reproduction (isolated env, no creds, fresh git repo, HEAD `a35ee9a0`): `timeout 5 env -i HOME=$ISOLATED_HOME PATH=$PATH CLAW_CONFIG_HOME=$PROBE/.claw-cfg claw prompt hello --output-format json > stdout.txt 2> stderr.txt` → stdout = **0 bytes**, stderr = 195 bytes containing `{"error":"missing Anthropic credentials…","exit_code":1,"hint":null,"kind":"missing_credentials","type":"error"}`, exit code 1. Confirms Gaebal's `1505208553793781792` pinpoint that `prompt` timeout + zero bytes was the prior state — HEAD `a35ee9a0` now correctly exits 1 with `kind:"missing_credentials"` **but the envelope is still routed to stderr** (issue #447 class, same class as prior entries #422, #435). **Contrast with `doctor`:** `claw doctor --output-format json 2>/dev/null` succeeds to stdout with `checks[auth].status:"warn"`, `api_key_present:false`, `auth_token_present:false` — but the auth check has no `prompt_ready:false` field. Automation that gates on `doctor` before invoking `prompt` must re-derive readiness from `api_key_present && auth_token_present` — there is no single canonical boolean. **Three compound problems:** (a) **stdout-empty on `--output-format json` failure**: same class as #447; `prompt`'s error envelope goes to stderr, not stdout. The canonical automation idiom `if ! result=$(claw prompt "q" --output-format json); then echo "$result" | jq .kind; fi` sees `$result=""` on failure — the jq call gets nothing. All `--output-format json` error paths must route JSON to stdout per #447 contract; (b) **`doctor` missing `prompt_ready` field**: `doctor --output-format json` already knows auth is absent (`api_key_present:false`) but surfaces no derived `prompt_ready:bool` or `prompt_blocked_reason:string` field. Automation must infer readiness from `api_key_present || auth_token_present || legacy_*_present` — a 5-field OR across legacy fields that is fragile as auth mechanisms evolve. A single `prompt_ready:false` (with `prompt_blocked_reason:"auth_missing"`) inside the `auth` check would give downstream a stable contract; (c) **`claw prompt` with no auth does no preflight and fires straight at the API**: the preflight check that `doctor` runs (auth discovery) is not reused by `prompt` to emit a fast typed error before attempting the network call. Both Gaebal's pinpoint (prompt hanging silently on older HEAD) and the current behavior (prompt hitting auth gate after a brief API attempt) stem from the same root: prompt does not short-circuit at the point where `doctor` already knows auth is absent. If `doctor` can emit `kind:"doctor"` with `auth.status:"warn"` in ~20ms without a network call, `prompt` should emit `kind:"missing_credentials"` in the same window and output it to stdout. **Required fix shape:** (a) `prompt --output-format json` must write the `kind:"missing_credentials"` JSON envelope to **stdout**, not stderr — same fix as #447 for all error envelopes; (b) add `prompt_ready:bool` and `prompt_blocked_reason:string|null` to the `auth` check in `doctor --output-format json`; derive it as `api_key_present || auth_token_present || legacy_saved_oauth_present`; (c) `prompt` must run the credential preflight check (same codepath as doctor's auth check) before attempting any API call and emit `{"kind":"missing_credentials","prompt_blocked_reason":"auth_missing"}` on **stdout** with exit 1 if the check fails; (d) `--output-format json` stdout routing fix must cover: `prompt`, `session list` (cross-ref #449), `skills uninstall` (cross-ref #431), `resume` (cross-ref #435), `acp serve` (cross-ref #443) — the full `kind:"missing_credentials"` class; (e) regression test: `claw prompt hello --output-format json` with no creds writes JSON to stdout (0 bytes stderr), exits 1, `kind:"missing_credentials"`, in under 200ms (no network attempt). **Why this matters:** `prompt` is the primary consumer entry point. Auth-absent failure routing to stderr breaks every automation wrapper that captures `$(claw prompt ... --output-format json)`. The `doctor` preflight metadata gap means auth-readiness checks require parsing 5 legacy fields instead of reading one boolean. Cross-references #447 (all JSON error envelopes on stderr), #449 (session list hits auth gate), #431 (skills uninstall hits auth gate), #357 (auth gate on local ops cluster), #422 (exit-code parity). Source: Jobdori live dogfood, `a35ee9a0`, 2026-05-16.
+
+464. **`--output-format` value parsing is strict-equal-only on `"text"`/`"json"` with five compounding gaps in `CliOutputFormat::parse` (`main.rs:600-611`): (1) **case-hostile** — `JSON`/`Json` rejected even though every UNIX enum-flag convention accepts both; (2) **whitespace-hostile** — `'json '` and `' json'` rejected, a known shell-interpolation footgun; (3) **no "Did you mean?" suggestion** for obvious near-matches like `JSON`/`json5`/`Json`; (4) **error format is always text-mode prose** — even when the operator is explicitly requesting JSON output (`--output-format=xml` → user clearly wanted machine-readable, gets text-prose error to stderr, chicken-and-egg bootstrap problem); (5) **classifier orphan** — the sentinel `"unsupported value for --output-format: ..."` is not registered in `classify_error_kind` so JSON-mode invocations would land with `kind: "unknown"` if it could ever reach that path. This is the second confirmed instance of the `#37×#77 classifier-orphan pattern` documented in #463 — features added without taxonomy registration silently degrade envelope quality** — dogfooded 2026-05-24 for the 14:30 Clawhip pinpoint nudge at message `1508114879985356992`, reproduced on local `./rust/target/debug/claw` `git_sha 003b739d` (origin/main `f8e1bb72`). Live matrix from clean isolated env (`env -i HOME=/tmp/iso21/home PATH=/usr/bin:/bin TERM=dumb`):
+
+    ```bash
+    $ for fmt in "xml" "JSON" "Json" "" "json5" "yaml" "json " " json" "JSON-LD"; do
+        printf -- '--- --output-format=%q ---\n' "$fmt"
+        claw --output-format "$fmt" status </dev/null 2>&1
+        echo "exit=$?"
+      done
+    ```
+
+    Every one of those 9 inputs produces identical-shape text-mode stderr prose:
+
+    ```
+    [error-kind: unknown]
+    error: unsupported value for --output-format: <VALUE> (expected text or json)
+
+    Run `claw --help` for usage.
+    ```
+
+    Note `[error-kind: unknown]` literally leaks into the user-facing prose — a fragment of the classifier debug output, not a friendly hint. The same `[error-kind: unknown]` prefix shows up on `claw login` per #463 — confirming the orphan-sentinel pattern crosses surfaces.
+
+    **Per-input failure-mode breakdown:**
+
+    | Input | Why it's a real bug class |
+    |---|---|
+    | `JSON` | UNIX convention: enum flags are case-insensitive (`curl --request POST`/`post`; `git --color always`/`ALWAYS`; `cargo --color=always`/`Always`). Claw is in a tiny minority of case-hostile parsers. |
+    | `Json` | Same as above; mixed-case is common in TitleCase environments. |
+    | `'json '` / `' json'` | Standard shell-interpolation footgun: `OUTPUT_FORMAT=$(some-cmd)` may include trailing newline; `--output-format "$OUTPUT_FORMAT"` then injects whitespace. `.trim()` is one line of defense. |
+    | `''` | Empty value from unset env var: `--output-format "$UNSET_VAR"`. Specialized error ("empty value supplied for --output-format") would clarify. |
+    | `json5` / `yaml` / `xml` / `JSON-LD` | No "Did you mean: json?" suggestion. Slash commands have this; subcommand parser has this (now); enum-value parser does not. Asymmetric quality. |
+
+    **Root cause traced:** `rust/crates/rusty-claude-cli/src/main.rs:600-611`:
+
+    ```rust
+    impl CliOutputFormat {
+        fn parse(value: &str) -> Result<Self, String> {
+            match value {
+                "text" => Ok(Self::Text),
+                "json" => Ok(Self::Json),
+                other => Err(format!(
+                    "unsupported value for --output-format: {other} (expected text or json)"
+                )),
+            }
+        }
+    }
+    ```
+
+    Five obvious enrichments missing: (a) `value.trim()` before match; (b) `value.to_lowercase()` before match; (c) levenshtein/prefix near-match suggestion; (d) specialized empty-value branch; (e) explicit `kind` discrimination so downstream classifier doesn't have to substring-match.
+
+    **The chicken-and-egg sub-bug (most severe):** A claw running `claw --output-format json status` to get structured output gets text-prose to stderr when the flag value is invalid. The very flag whose purpose is to request JSON cannot itself produce a JSON error envelope. The minimum fix is: if `--output-format` parsing fails AND any earlier argv contained `--output-format=json` OR `--output-format` with `json` as next arg (i.e., the operator's intent is clearly JSON), THEN emit the parse error as a JSON envelope on stderr instead of text prose. Bootstrap problem solved.
+
+    **Why distinct from existing items:** Surfaces a different facet of the broader **classifier-orphan / sentinel-not-typed pattern** documented in #463 (login/logout) and #455 (missing_credentials hint shape) and #449 (prompt JSON error routed to stderr). #463 covered the SAME root function (`classify_error_kind`) for a DIFFERENT sentinel ("has been removed"). This entry covers a DIFFERENT sentinel ("unsupported value for --output-format") with the SAME root, plus FOUR additional concerns specific to enum-value parsing (case, whitespace, near-match, empty). #109 covers config-loader warnings flattened to stderr prose (loader-side, not CLI-flag-side). #108 covers subcommand typos with no "Did you mean?" — analogous design pattern but different parser layer (subcommand name vs flag-value). **None** of these existing entries document the `CliOutputFormat::parse` strict-match-with-five-gaps surface. **Why this matters:** (1) **CI/automation reliability** — every CI step that constructs `--output-format $VAR` with VAR from another tool's output (Make, Justfile, shell completion, env file) is one trailing-newline or one casing mismatch away from text-prose-to-stderr instead of structured JSON. (2) **Bootstrap chicken-and-egg** is the worst flavor: the flag exists specifically to enable machine parsing, and the flag's own error path bypasses that machine parsing. (3) **Cross-surface asymmetry erodes trust** — slash commands suggest, subcommand parser now suggests (per #108 fix), but flag-enum parser doesn't. A claw learning the CLI grammar by trial-and-error gets contradictory teaching about the quality of suggestions. (4) **Same classifier-orphan pattern as #463** — `[error-kind: unknown]` text leakage proves the classifier-string roundtrip is happening even on simple flag-parse failures. The taxonomy gap from #77 is structurally chronic; per-instance fixes alone won't resolve it. (5) **Discovery method itself is a signal** — a 30-second hypothesis sweep covering 9 inputs caught 5 distinct bug classes. Strict-match parsers without trim+case+near-match+empty handling have a high latent-bug density per LoC. The classifier should ship as a typed result with structured envelope from day one. **Required fix shape:** (a) **`CliOutputFormat::parse` enrichments:** trim input; lowercase before match; explicit empty-value branch with specialized error; levenshtein-based "Did you mean?" suggestion (reuse the slash-command suggester from `commands::resolve_skill_invocation` or the subcommand suggester being added for #108). (b) **Bootstrap fix:** when `--output-format` parsing fails AND the operator's argv contains the literal token `json` adjacent to `--output-format`, emit the parse error as a JSON envelope to stderr so JSON-requesting consumers can decode it. (c) **Classifier registration:** add `"output_format_invalid"` to `classify_error_kind`, keyed on `"unsupported value for --output-format"`. (d) **Strip `[error-kind: ...]` debug prefix from text-mode user-facing output** — it's classifier internals leaking into the UX (also fixes #463's same complaint). (e) **Regression coverage in `output_format_contract.rs`:** parameterized test asserting 9 invalid inputs each produce the expected enriched behavior (case-normalized acceptance, trimmed acceptance, near-match suggestion, empty-value error, JSON-envelope bootstrap output when intent is JSON). **Acceptance check (one-liner):** `claw --output-format JSON status 2>&1 | jq -e '.kind == "output_format_invalid"'` should print `true` for the bootstrap-fix path; `claw --output-format JSON status; echo $?` should exit 0 after the case-normalization fix; `claw --output-format json5 status 2>&1 | grep -q "Did you mean: json"` should pass after the suggester fix. Source: gaebal-gajae dogfood for the 2026-05-24 14:30 Clawhip pinpoint nudge at message `1508114879985356992`. Pre-grep gate filtered 9 hypotheses → 2 fresh (W=`--output-format` value handling, Y=base URL validation); W selected because failures reproduce in credential-free path with multiple distinct sub-issues. Y deferred to a future tick because it requires actual HTTP-send path to surface failure, and the BASE_URL pinpoint is more about the `status`/`doctor` envelope failing to surface "configured to talk to malformed URL" warnings — different fix shape, different code path. Coordination note: Jobdori took #469 (`/compact` slash divergence) this tick and acknowledged F (CLAW_CONFIG_HOME validation, 5-mode silent failure surfaced in #463 tick) as "next confirmed but unfiled," so F is intentionally NOT covered here to avoid duplicate filing.