Root-cause fix for the 'TTS voice bubble broken' issue family (#57048,
#54589, #57213, #58845, #14841, #45557, #57049). Two class-level defects:
1. Several backends silently write MP3/WAV bytes into a .ogg output path
(Edge only emits MP3, Piper writes WAV, xAI writes MP3, some
OpenAI-compatible servers ignore response_format=opus). Platforms that
need real Ogg/Opus render 0-second/broken voice bubbles. Instead of
per-provider patches, text_to_speech_tool now sniffs magic bytes once
after synthesis (_sniff_audio_container) and repairs the container
centrally (_repair_ogg_container): ffmpeg transcode in place, or rename
to the honest extension when ffmpeg is unavailable. Covers every
current and future provider, including command providers and plugins.
2. want_opus only recognized Telegram, so Matrix/Feishu/WhatsApp/Signal
auto-TTS voice replies were synthesized as MP3 and delivered as broken
attachments. New OPUS_VOICE_PLATFORMS set covers all voice-bubble
platforms.
_convert_to_opus refactored onto a shared _ffmpeg_transcode_to_opus that
supports safe in-place transcodes (-f ogg forced muxer, temp file +
os.replace).
Tests: tests/tools/test_tts_container_repair.py (13 tests incl. a live
ffmpeg round-trip); full TTS suite green (153 + 187 passed).
Class-level fix for the 'STT transcribes the wrong language' issue family
(#55551, #50181 and siblings). Previously language handling was per-provider
chaos: local honoured stt.local.language, Groq/OpenAI/Mistral/DeepInfra sent
no language hint at all, xAI silently forced 'en', ElevenLabs used its own
language_code key, and there was no global setting.
- New _resolve_stt_language() helper: stt.<provider>.language >
stt.language (new global key) > HERMES_LOCAL_STT_LANGUAGE > auto-detect.
- Threaded through ALL providers: local, local_command, groq, openai,
mistral, xai, elevenlabs, deepinfra (shared OpenAI handler), command
providers, and plugin dispatch.
- xAI no longer forces English when nothing is configured (auto-detect).
- Mistral Voxtral now receives a language hint when configured.
- stt.groq.model is now honoured from config (previously env-only).
- DEFAULT_CONFIG gains stt.language, stt.groq, stt.xai, stt.mistral.language.
- Tests: tests/tools/test_stt_language_resolution.py (11 tests, sabotage-
verified) + full transcription suite green (236 passed).
Builds on cherry-picked contributor work from #19786 (@zombopanda),
#23161 (@materemias), #50684 (@BlackishGreen33).
`stt.groq: null` in config.yaml yields `groq_cfg = None`, so the
subsequent `.get("language")` raised AttributeError. Use `or {}`
(matching main's widened xai/local provider guards) and add a
`{"groq": None}` regression test confirming auto-detect stays intact.
Addresses hermes-sweeper review on #23161.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019Zc7Tgei4kav4n6WSsBq5F
- Normalize stt.groq.language: cast to str, strip, treat
empty/whitespace as unset (parity with xAI's str().strip()).
- Clarify "blank = auto-detect" inline comments in 4 docs/configs to
reflect the env-var fallback (HERMES_LOCAL_STT_LANGUAGE).
- Document that HERMES_LOCAL_STT_LANGUAGE also drives the local
faster-whisper provider, not just the CLI fallback.
- Add unit tests covering: omitted language when unset, config-supplied
language, env fallback, config-over-env precedence, whitespace
normalized to unset.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Read stt.groq.language from config.yaml (with HERMES_LOCAL_STT_LANGUAGE
env fallback) and forward it to the Groq Whisper API to skip
auto-detection on known-language audio. Omit when unset so Groq
auto-detects, preserving today's behavior. Bonus: swap xAI's hardcoded
env literal for the LOCAL_STT_LANGUAGE_ENV constant for consistency.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add optional stt.openai.language config for OpenAI transcription. Forward non-empty hints to the API while preserving auto-detection when unset. Document the config-only setting, add its default, and cover configured and unset request arguments.
Direct tool HTTP calls already identified as Hermes-Agent, but the main
OpenAI-SDK chat path still sent OpenAI/Python. Set Hermes-Agent/<ver> for
api.x.ai clients (xai + xai-oauth) so normal text traffic is attributed correctly.
Tool-schema assembly at CLI/Desktop startup runs the browser-family
check_fns (browser, browser_cdp, browser_dialog, browser_vision). Each
of those gates called _get_cdp_override(), which resolves the configured
endpoint over HTTP (GET /json/version, timeout=10) — so a *stale*
browser.cdp_url pointing at a dead debug browser cost ~7 serial blocking
socket connects before the banner rendered. Measured on a real Windows
install with a dead http://[::1]:9222 config: 15.1s of an 18s launch,
with no warning or error — just mystery slowness. The value is easy to
leave behind: /browser connect writes a session-scoped env override, but
'hermes config set browser.cdp_url' persists forever while the debug
Chrome it pointed at dies on the next browser restart.
Split the helper:
- _get_cdp_override_raw() — returns the configured value (env var or
config.yaml) with zero network I/O. Used by every is-it-configured
gate: check_browser_requirements, _browser_cdp_check, _is_local_mode,
_is_local_backend, _navigation_session_key, _should_inject_engine
(via _is_local_mode), and the hermes doctor chromium-skip check.
- _get_cdp_override() — unchanged contract (raw + /json/version
resolution), now only called on paths that are about to connect:
session creation and the dialog-supervisor attach.
This follows the existing rule in check_browser_requirements ('do not
execute agent-browser --version here') and the browser.manage status
path, which already banned _get_cdp_override for exactly this reason
(test_browser_manage_status_does_not_call_get_cdp_override): schema
assembly must not perform blocking I/O.
A/B on the same machine, same dead endpoint: get_tool_definitions()
15.08s unpatched -> 1.89s patched, with browser_cdp/browser_dialog
still advertised (gate now keys off configuration, not reachability —
matching the documented lazy-supervisor contract in
_browser_dialog_check).
Adds a regression test asserting the browser_cdp check_fn never touches
the network.
Port from nearai/ironclaw#5149 (the describe-first live-hardening fix in
their progressive tool disclosure work): when a model invokes a deferred
tool through the tool_call bridge without the schema-required arguments,
return the tool's parameter schema instead of dispatching blind.
Pre-fix, a blind call produced an opaque downstream failure
("[TOOL_ERROR] Tool execution failed: KeyError: 'document_id'") that
teaches the model nothing about what the tool expects — IronClaw observed
cheap models looping ~30 identical invalid calls until the iteration
budget died. Post-fix, the model repairs the call in one round-trip.
- tools/tool_search.py: new validate_deferred_call_args() — key-absence
check of schema 'required' fields only; no type checking (coerce_tool_args
already repairs types downstream); fails open on any validator error so
it can never block a legitimate dispatch.
- model_tools.py: probe after the scope gate in the bridge dispatch.
- agent/tool_executor.py: probe in both unwrap sites (concurrent +
sequential) before the underlying tool replaces the bridge; sequential
path flattens the payload to match its {"error": str} wrapping.
- tests: TestDeferredCallSchemaProbe — blind call returns schema (not
KeyError), valid/optional calls dispatch, unvalidatable tools fail open,
out-of-scope rejection unchanged.
Port from openai/codex#33464: GNU rm permutes options, so
`rm build/ -rf`, `rm build/ -r -f`, and `rm build/ --recursive
--force` are equivalent to the flags-first spellings — but every
existing rm pattern required the flag group BEFORE the path, so these
spellings ran with no approval prompt at all (proven live on main).
The hardline floor was NOT affected: protected paths (/, system dirs,
$HOME) match regardless of flag position because the hardline path
matcher does not require flags. The gap was the approval-prompt layer
for arbitrary paths.
New DANGEROUS_PATTERNS entry with a tempered operand run: cannot cross
command separators (; | & newline), quotes, or a bare -- end-of-options
separator (after --, -rf is a literal filename). Flag token must be
whitespace-anchored so the r inside long options like --registry does
not count.
9 positive + 7 negative shapes in tests; approval cluster (868 tests)
green.
Port from anomalyco/opencode#38133/#38134 (patch Unicode matching corpus):
extend UNICODE_MAP with the Zs space-separator family (en/em quad, en/em/
three-per-em/four-per-em/six-per-em/figure/punctuation/thin/hair spaces,
narrow NBSP, medium mathematical space, CJK ideographic space) and the
Unicode minus sign U+2212.
Before: a file containing typographic spacing (French narrow NBSP, CJK
ideographic spaces, math minus) never matched a model's ASCII old_string
via the precise strategies — the edit only succeeded through the
similarity-based context_aware fallback, which (a) can pick the wrong
region (#54572 family) and (b) silently flattens the file's Unicode to
ASCII on replacement. After: these match at unicode_normalized (strategy
7), whose _preserve_unicode_in_replacement keeps the file's typographic
characters in unchanged spans.
All additions are 1:1 mappings, so the existing position-mapping and
preservation logic apply unchanged. Proven live before/after with a
multi-line probe; 59 fuzzy-match + 188 file-tools/patch/skill-manager
tests pass.
Inspired by Claude Code 2.1.214, which added permission prompts for
container-CLI commands (including the Podman shim) carrying
daemon-redirect flags (--url, --connection, --identity, remote mode)
that previously ran without one.
A daemon redirect makes a local-looking command operate on a different
(often remote) daemon, silently acting on production infrastructure.
Any container-CLI invocation carrying a redirect now requires approval
regardless of subcommand:
- -H/--host and --context global flags (value required, global-flag
position only — bare -h help and run-level -h <hostname> stay allowed)
- context use (persistently switches the default daemon)
- podman --url/--connection/--identity and -r/--remote
- DOCKER_HOST=/DOCKER_CONTEXT=/CONTAINER_HOST=/CONTAINER_CONNECTION=
environment prefixes
Sibling-site widening: the existing container lifecycle rules matched
only the verb directly adjacent to the binary name, so a global flag or
a compose -f file flag slipped past the guard, and the legacy hyphenated
compose binary was never covered. They now tolerate global flags — the
same treatment the 'hermes ... gateway' rule already has — and match the
hyphenated compose binary.
Validation: 33 new tests; 339 pass in test_approval.py + new file;
442 pass across the adjacent guard suites; E2E battery of 12 dangerous
+ 17 safe commands via real imports, hot path ~330us/call.
Root cause: delegate_task children run through three nested daemon-thread
layers (async-delegation executor -> per-child timeout executor -> the
interrupt worker interruptible_api_call spawns). After multi-day gateway
uptime the deepest layer wedges BEFORE the socket opens — the same
fingerprint as the gateway-cron hang (#62151): zero stale-detector output
(the worker never reaches dispatch), all providers, foreground/restart
works. The cron fix (should_use_direct_api_call) explicitly excluded
delegation 'for lack of evidence' — #60203 is that evidence.
- should_use_direct_api_call: extend the inline gate to delegated
children, detected via the delegation ContextVar set by
_run_single_child (platform='subagent' stamp as fallback). Scope
unchanged otherwise: chat_completions wire only; Codex/Anthropic/
Bedrock/MoA keep their established workers. Interrupts still work —
the inline path registers _active_request_abort, which interrupt()
invokes cross-thread (same mechanism the #72227 stall monitor uses).
- _dump_subagent_timeout_diagnostic: dump ALL thread stacks (bounded,
40), not just the conversation worker — a pre-HTTP wedge is
indistinguishable from a slow provider without seeing where the
nested helper threads sit.
Line-based grep on export -p only drops the first declare -x line, so a
newline in HERMES_SESSION_CHAT_NAME/USER_NAME leaves shell payload in the
shared snapshot and runs it on the next source. Unset bridged vars in a
subshell before export -p instead (issue #71296).
`hermes skills update` could silently replace a same-named skill from a
DIFFERENT registry — deleting the user's files and rewriting the lockfile's
recorded provenance. Two defects on main:
- tools/skills_hub.py: check_for_skill_updates fell back to *all* sources
(`... or sources`) when no adapter matched the recorded source, so any
registry with a same-named skill could satisfy the fetch and be reported
as an update — silently reassigning provenance. Now reports the entry as
`unavailable` instead of cross-registry fallback.
- hermes_cli/skills_hub.py: do_update called do_install with a bare, slash-
less identifier and no source constraint, letting _resolve_short_name fuzzy-
match a same-named skill in another registry. do_install now takes an
optional source_id pin that ABORTS (rather than falls back) when no adapter
matches, and do_update forwards the lockfile's recorded source as that pin.
Includes regression tests reproducing the cross-registry hijack.
Salvaged from PR #72216 onto current main; re-attributed to the human
contributor.
Co-authored-by: menhguin <menhguin@users.noreply.github.com>
Widen the cherry-picked /diff base (#4839 by @SHL0MS) into one
cross-surface implementation, folding in the review feedback and the
best ideas from the two sibling PRs (#22703, #53527):
- tools/working_diff.py: shared git collection layer — unstaged
(default), staged, and all (vs HEAD) modes; untracked files folded in
via `git diff --no-index` so new files appear as additions (Codex
/diff parity); shlex-split arguments preserve quoted paths.
- CLI: handler moved to hermes_cli/cli_commands_mixin.py per the
current god-file decomposition (dispatch stays in cli.py), renders
through the rich console with a 400-line terminal-flood guard.
- Gateway: _handle_diff_command in gateway/slash_commands.py + dispatch
in gateway/run.py; fenced ```diff output truncated to 60 lines /
3000 chars before the platform senders apply their own per-platform
message clamps (tool-progress-style layered truncation). Localized
strings in all 17 locale catalogs.
- /diff session (from #53527): cumulative checkpoint-baseline diff of
everything Hermes changed, via new CheckpointManager.session_diff();
docstring records the retained-baseline approximation caveat from
review. Works on both surfaces; degrades with an actionable message
when checkpoints are off.
- Slack: /diff routed via /hermes diff (50-slash cap; keeps
telegram-parity test green and /version native).
- Registry: cross-surface CommandDef with staged|all|session
subcommands; docs: slash-commands reference (CLI + gateway tables +
both-surfaces list) and hermes-agent skill reference.
- Tests: tests/tools/test_working_diff.py (real git repos),
tests/hermes_cli/test_diff_command.py (real git + stubbed checkpoint
manager), tests/gateway/test_diff_command.py (end-to-end handler,
real checkpoint store), TestSessionDiff in
tests/tools/test_checkpoint_manager.py.
Salvaged from the /diff PR cluster #4839 + #22703 + #53527.
Co-authored-by: Ninso112 <ninso112@proton.me>
Co-authored-by: Harshkamdar67 <harshkamdar67@gmail.com>
After N consecutive guardian denials in a session the deny message escalates to a hard-stop instruction. Inspired by ChatGPT Work auto-review circuit breaker.
Cross-surface coverage #23768 missed (per review feedback):
- tools/clarify_gateway.py: _ClarifyEntry carries a multi_select flag
(register() accepts it; signature() exposes it to adapters).
_coerce_text_response now parses multi-select replies — comma- or
space-separated numbers ('1,3' / '1 3'), exact labels, dedup — into a
JSON array string that _parse_multi_select_response decodes into a
list. Out-of-range/unknown tokens reject the reply (native button UI)
or fall back to custom text (awaiting_text/'Other' mode).
- gateway/run.py: _clarify_callback_sync accepts multi_select and
registers it on the pending entry.
- gateway/platforms/base.py: default numbered-list text fallback tells
the user multiple selections are allowed and how to reply.
- tui_gateway/server.py: clarify_callback passes multi_select through
the clarify.request payload as a hint; renderers without checkbox
support ignore the field and remain single-select-compatible.
- tests: 13 new gateway tests (flag storage, comma/space/single-number
parsing, label matching, out-of-range rejection, dedup, end-to-end
resolve, single-select regressions).
Follow-ups to the salvaged #23768 commit, which targeted a pre-79559214
codebase:
- agent/tool_executor.py + agent/agent_runtime_helpers.py: pass
multi_select at both current clarify dispatch points (the PR's
run_agent.py edits landed on dead code paths).
- tools/clarify_tool.py: replace the broad TypeError-retry in
_invoke_callback with inspect.signature detection, so a compatible
callback that raises TypeError internally is not invoked twice
(addresses hermes-sweeper review feedback on #23768).
- tests: cover single-invocation on internal TypeError, legacy 2-arg
callbacks, **kwargs callbacks, and registry handler multi_select
pass-through (schema arg → handler → callback).
hermes update's lazy-refresh pass re-asserts LAZY_DEPS pins whenever the
package is present (active_features() is presence-based). The
tool.trace_upload pin huggingface-hub==1.2.3 sat below transformers'
>=1.5.0,<2 requirement, so every update force-downgraded the shared
package and broke Hindsight local embeddings on daemon startup (#60783).
Keep the exact-pin security posture — no ranges — but move the pin to
1.24.0 (current) and bump uv.lock in lockstep (uv lock --upgrade-package
huggingface-hub: hub 1.4.1->1.24.0, hf-xet 1.3.1->1.5.2, click
8.3.1->8.4.2, drops typer-slim), so the entire tree converges on ONE hub
version. The refresh pass now reports 'current' with zero churn.
Invariant tests (not snapshots): the lazy pin must equal the uv.lock
resolved version, and must sit inside transformers' accepted window.
HfApi surface used by trace upload (whoami/create_repo/upload_file)
verified present with identical kwargs on 1.24.0 in a live venv.
Completes #51690 on top of the salvaged #60378 timeout metadata:
- async_delegation: terminal 'stalled' events now carry structured
stall context (stalled_after_quiet_seconds, stall_threshold_seconds,
stall_phase idle|in_tool, stall_grace_seconds) on both single and
batch paths, persisted in the durable row so restart-restored events
keep it. Mirrors the sync path's timeout_seconds/timed_out_after_
seconds/timeout_phase from #60378.
- list_async_delegations(): exposes seconds_since_progress and live
children_activity (per-child api_calls, current_tool,
seconds_since_activity) sampled from the dispatch's progress_fn
outside the records lock; private monitor bookkeeping and callables
never leak.
- /agents (CLI + gateway): background delegations render per-child
activity rows, quiet-time hints, and the stalling state; gateway
section is new (previously async delegations were invisible there).
New locale key gateway.agents.background_delegations in all 17
catalogs.
Tests: stall-metadata event shape, live-listing projection, gateway
/agents rendering (real registry dispatch, sabotage-verified), sync
timeout metadata fields, non-timeout None contract.
Add timeout_seconds, timed_out_after_seconds, and timeout_phase to timeout
results so parent agents and users can distinguish timeouts before the
first LLM call from timeouts after one or more API calls.
Also attach diagnostic_path to the N>0 API-call timeout error message,
matching the existing zero-API-call timeout path.
Addresses part of #51690 and #17308.
Include last_activity_ts in the progress token sampled from each child.
_touch_activity ticks on every streamed chunk ('receiving stream
response'), every tool transition, and API-call start/completion — so a
child mid-stream on a long response is alive even though api_call_count
only advances when the call completes. Same liveness signal as the
compaction inactivity budget (PR #71508): if tokens are flowing it never
dies; staleness is measured from the last streamed token / tool activity
/ API call.
Replace the wall-clock timeout watchdog (from #60234) with progress-based
staleness detection, on by default with zero config:
- The async registry now accepts a progress_fn per dispatch; delegate_task
wires a sampler over the batch's child agents (api_call_count +
current_tool from get_activity_summary()).
- A single monitor thread sweeps running delegations: a child whose
progress token keeps advancing is never touched, no matter how long it
runs. A frozen token past the stale threshold (450s idle / 1200s
in-tool, mirroring the sync-path heartbeat monitor) marks the record
'stalling' and interrupts the child.
- A stalling child that unwinds within the grace window (120s) finalizes
through the NORMAL path, preserving its partial results. One that never
returns is force-finalized with a terminal 'stalled' completion event so
the owning session hears an outcome and the async slot frees.
- Late runner returns after force-finalization are deduped by the
begin/push/finish finalization split (kept from #60234).
Why not a timeout: delegation.child_timeout_seconds defaults to 0 by
deliberate design (DEFAULT_CHILD_TIMEOUT rationale) — a timeout-based
watchdog never arms for default configs, leaving the reported silent-
profile symptom (#60203) unfixed, and when armed it kills legitimately
slow heavy subagents mid-task. Progress detection distinguishes 'wedged
at first API call' from 'grinding through a 2h review'.
Builds on izumi0uu's finalization-atomicity work from #60234.
Async background delegation can leave gateway sessions holding only a dispatched handle when the detached runner wedges before it can return and enqueue a completion. Enforce the configured child timeout in the async registry so the parent observes a terminal timeout event and the async slot is released.
Constraint: Issue #60203 reports long-lived gateway processes with background child delegates that never produce completion events despite child_timeout_seconds being configured.
Rejected: Relying only on _run_single_child timeout handling | it cannot finalize the async registry when the outer runner thread itself never reaches normal completion.
Confidence: high
Scope-risk: narrow
Directive: Keep background delegation completion owned by the async registry whenever detached workers can outlive the caller's immediate control.
Tested: .venv/bin/python -m pytest tests/tools/test_async_delegation.py tests/tools/test_delegate_subagent_timeout_diagnostic.py tests/tools/test_delegate.py -q
Tested: .venv/bin/python -m ruff check tools/async_delegation.py tools/delegate_tool.py tests/tools/test_async_delegation.py
Tested: git diff --check
Not-tested: Multi-day real gateway degradation; covered with deterministic stuck-runner registry tests.
'Refreshing cua-driver (Computer Use)...' could hang for minutes on
Windows: when the driver's native check-update verb returned an
indeterminate result (old driver without the verb, offline, GitHub
rate-limited, or the probe timing out), install_cua_driver(upgrade=True)
fell through to the full upstream installer — a silent, output-captured
run with a 660s ceiling, plus install.ps1's 600s concurrency-lock wait
on Windows on top. Every 'hermes update' paid that cost.
Two changes:
- install_cua_driver() grows require_confirmed_update: with it set, an
indeterminate check keeps the installed version and returns fast,
printing the force path (hermes computer-use install --upgrade).
'hermes update' passes it; the explicit --upgrade CLI keeps the old
fall-through so a force refresh still works when the check can't
answer.
- cua_driver_update_check() default timeout is now 25s on Windows
(8s unchanged on POSIX): first-spawn of the exe under Defender /
SmartScreen routinely exceeds 8s, and a false timeout is exactly the
indeterminate result that used to trigger the multi-minute reinstall.
MEDIA_DELIVERY_EXTS in gateway/platforms/base.py omitted .3gp, causing
MEDIA: tags with .3gp files to leak as plain text instead of being
extracted for native video delivery. _VIDEO_EXTS in
tools/send_message_tool.py and _MIGRATION_VIDEO_EXTS in the Feishu
adapter omitted .webm, causing .webm files to be classified as
documents instead of video on Telegram and other platforms.
Both extensions are already present in every gateway-side _VIDEO_EXTS
definition (run.py, kanban_watchers.py, weixin.py, base.py local).
Closes#71621, Closes#71603
The include/exclude filter matched exact names only — glob-style entries
('*_radar_*') silently matched nothing, so a Cloudflare flat-mode config
meant to trim 3,320 tools to ~1,900 actually registered 3,319. Unmatched
patterns produced no warning.
- matches_name_filter(): exact membership first (O(1) for literal lists),
then fnmatch.fnmatchcase for entries containing * ? [ — same pattern
semantics as approvals.deny. Entries without metacharacters stay
strictly literal ('docs' never matches 'docs_search').
- _should_register() uses it for both include and exclude (symmetric)
- hermes mcp tools picker (mcp_config.py) pre-selection uses the same
matcher so the UI agrees with runtime registration
E2E against the live Cloudflare capture with the real exclude list:
3,320 -> 1,905 surviving (1,415 excluded); radar/DLP gone,
purge_cache/dns_records kept. 220/220 mcp_tool tests (4 new).
Teknium review changes on the tiered policy:
1. threshold_pct default 10 -> 5 (listing budget = min(5% of context,
listing_max_tokens)); unknown-context fallback 20K -> 10K.
2. Tier 2 no longer leaves the model blind: when even names-only doesn't
fit, the bridge description now carries a one-line-per-server summary
('cloudflare (3320 tools)') plus an instruction to search FIRST rather
than substitute a generic tool or claim the capability is missing —
the measured tier-2 failure mode (core-tool substitution) at zero
meaningful token cost (~50 tokens/server).
3. Listing degradation is now PER SERVER, largest first: one oversized
server (Cloudflare) collapses to its summary line while small
co-attached servers (Linear) keep their full per-tool listings
('mixed' form). Previously global: attaching Cloudflare next to
Linear silently cost Linear its listing. Greedy fit is deterministic
(size then label) so the rendered block stays byte-stable per catalog
— prompt-prefix cache safe.
E2E on real captures (defaults, 200K ctx): linear alone -> tier 1 full;
unreal alone -> tier 2 groups (5% budget) / tier 1 names at 1M;
cloudflare alone -> tier 2 groups; linear+cloudflare -> tier 1 MIXED
(linear fully listed, cloudflare summarized). 48/48 tests.
