`_segmented_music` mixed the video's audio with the shaped music via
`amix=inputs=2`, whose default `normalize=1` scales every input by 1/inputs
(x0.5, -6 dB). Unlike `_mix` and `_full_mix`, this path has no `loudnorm` stage
afterward to re-normalize, so the narration was permanently attenuated across
the entire timeline — including the stretches where the music volume expression
evaluates to 0. A one-second music segment quietly dropped the narration by
~6 dB for the whole video.
Add `normalize=0` to the amix: the music is already scaled to `music_volume`
by the `volume` expression, so speech passes at unity. Verified with ffmpeg —
narration in a no-music region tracks the stereo/aac conversion baseline
instead of sitting 6 dB below it.
The tool advertised `multiple_outputs: True`, accepted `n` (1-4) in its schema,
requested `n` images from the API, and scaled `estimate_cost` by `n` — but the
result handling was hardcoded to `response.data[0]`. Images 1..n-1 were decoded
never, written never, and absent from `artifacts`, so a caller who set `n=4`
paid for four images and received one.
Iterate over `response.data`, writing each image to a distinct path (suffixed
`_1`, `_2`, … when several are requested, mirroring `grok_image` /
`dashscope_image`), and return `outputs` / `images_generated` alongside the
full `artifacts` list. A single image keeps its exact requested path.
From an independent review of the branch:
- Regression: existing non-raster-but-showable visuals (.svg diagrams) were
dropped by the renderable filter — they were served fine before via <img>
(/thumb passes SVG through). Add .svg to MEDIA_IMAGE_EXT.
- Doc accuracy: the board dedupes decisions by (category, subject), not category
alone (category-only would wrongly merge distinct decisions that share a
category, e.g. TTS vs image provider_selection). Correct AGENT_GUIDE to say
the pair is the key and to reuse the same subject when re-logging.
- Coverage: the new visual-selection logic was untested (which let the earlier
missing-file regression through). Add TestStoryboardVisualSelection covering
the .tsx-animation exclusion, snapshot fallback (exact + <id>_* match), SVG
renderability, the preserved missing-file indicator, and takes = renderable.
Not changed (reviewed, deliberate): .narr clamp at --fs-scale 1.16 degrades
gracefully via the fade + click-to-expand modal; decision dedupe stays keyed on
(category, subject) as the more-correct behavior.
The prior dunder denylist was still bypassable via print.__self__ (the builtins
module) -> .open(...), reachable with no import and no bare open/__builtins__/
getattr name. Enumerating dangerous dunders is whack-a-mole, so block ALL
dunder attribute access generically and allow only the tiny set legitimate
scenes need (super().__init__, occasional Type.__name__). This closes the
print.__self__ / .__class__ / .__globals__ introspection-escape class at once.
Static analysis still has a ceiling — a real subprocess sandbox is the complete
fix — but the default path no longer executes the reported secret-read payloads.
Adds regression tests for print.__self__ and for super().__init__ staying allowed.
Refs #219
The scan only flagged dangerous builtins as direct call targets (ast.Name func)
and dunders as attribute access, so it missed indirection like
`__builtins__['open']('.env').read()` and `getattr(o, '__class__')` — the
default path still executed secret-reading code.
Block dangerous identifiers wherever they appear as a bare name (open, eval,
exec, compile, __import__, __builtins__, getattr/setattr/delattr, globals/
locals/vars) rather than only as a call target, and extend the blocked dunder
set (__class__, __dict__, __getattribute__, __reduce__, ...). This closes the
reported no-import bypass while genuine math scenes still pass.
Still defense-in-depth, not a full sandbox; the allow_unsafe_code opt-out and
explicit code-execution contract remain. A subprocess-level sandbox is the
right follow-up for complete isolation.
Refs #219
math_animate writes caller-supplied Python to scene.py and runs Manim on it —
arbitrary local code execution with no boundary surfaced in the tool contract.
In an agent-driven system the scene_code may be LLM-generated or influenced by
untrusted prompt content, so import-time code or construct() could read
secrets/SSH material, open network connections, or spawn subprocesses.
Add a static AST safety scan that rejects dangerous imports (os, subprocess,
socket, requests, ctypes, ...), dangerous builtins (eval/exec/compile/open/
__import__), and sandbox-escape dunders (__globals__, __subclasses__, ...)
before Manim runs. Genuine math scenes (manim, numpy, math, ...) pass
untouched. This is defense-in-depth, not a sandbox: a determined attacker can
evade a static denylist, so it is paired with an explicit allow_unsafe_code
opt-out and a tool contract (schema + side_effects) that names the boundary.
Closes#219
- F-01: cost bar crit (red) state past 90% of budget
- F-02: normalize() hardens fetched board state against sparse payloads
- F-03: /thumb 404s for videos with no extractable poster frame instead
of serving raw video bytes
- F-04: checkpoint artifact path refs only resolve inside the project dir
- F-05 (board half): stall detection — in_progress stage with no disk
activity >10min renders red 'stalled?' + header badge flips to STALLED?;
verified against the real wedged why-cities-glow project
- eval harness from dogfood session committed (visual regression +
interaction smoke, capture watcher, server/gate test suites) +
regression tests for each finding; 46 backlot tests green, visual eval
green (restage-before-capture note logged)
review_source_media deliberately returns files:[] with a 'no source media —
fully generated production' summary when no user media is supplied or none can
be reviewed, but the schema declared files.minItems:1, so that intended
artifact failed its own validation. Relax files.minItems to 0 to match the
code's deliberate empty-media state (planning_implications still carries an
entry, so its minItems:1 remains satisfied).
Adds a regression test validating the no-source-media artifact.
Closes#269
Check 2 flagged 'N consecutive same-size shots' from a count of every equal
adjacent pair across the whole plan, not the length of any real run. So three
separate 2-shot groups (wide,wide,cu,cu,med,med) tripped a false '3
consecutive' violation, while a genuine run of 3 (only 2 pairs) was never
flagged. Track the current run length, reset on change, and compare the longest
run >= 3.
Adds regression tests: non-consecutive pairs pass, a true run of 3 is flagged,
unspecified shots don't form a run.
Closes#268
full_mix with ducking enabled (the default) failed for a single narration
track + one music bed — the most common shape — because the ducking branch
appended an acopy[speech_dup] filter whose output pad was never consumed,
leaving the filtergraph with a dangling output that ffmpeg rejects.
For a single speech track speech_out is '[a0]' (starts with '[a'), so the
guarded append fired; the compensating pop() only removes the empty-string
case from the multi-speech branch, so the dead pad survived exactly in the
single-narration case. The speech stream is already re-derived for the final
mix via [speech_out], and ffmpeg auto-splits the reused input label, so the
duplicate is unnecessary. Multi-speech and SFX paths are unaffected.
Adds regression tests for single- and multi-narration full_mix with ducking.
Closes#265
The timeout handling only took effect on a direct _remotion_render() call. The
high-level execute(operation='render') path goes through _render(), which builds
a fresh remotion_inputs dict (edit_decisions, output_path, profile) and dropped
remotion_timeout_ms — so callers of the documented operation='render' path never
got the timeout passed to the Remotion CLI. Forward it there.
Adds a test exercising _render() (not just _remotion_render()) to cover the
high-level forwarding path.
Refs #217
The high-level Remotion render path hid the useful failure reason. run_command
runs with check=True + capture_output, so a non-zero exit raised
CalledProcessError whose str() is only 'returned non-zero exit status 1' — the
actual Remotion diagnostics in stderr were dropped. Catch CalledProcessError
and surface the stderr/stdout tail, and TimeoutExpired with an actionable hint.
Also add a creator-facing remotion_timeout_ms input, passed through as
Remotion's --timeout (headless-browser setup + delayRender). Slow browser
startup on restricted networks previously failed opaquely at the default 30s
with no way to raise it. The subprocess timeout is widened to match so
run_command does not kill Remotion before its own timeout fires.
Closes#217
- backlot/state.py: BoardState from disk — stage rail with gate audit
(gate_skipped detection from history/), scene_plan x script x
asset_manifest storyboard join, takes, generating-state from events,
media discovery incl. atelier root-render heuristic, degradation ladder,
library summaries
- backlot/server.py: FastAPI on 4750 — /api/projects, /api/project/{id}/state,
SSE change feeds (project + library) fed by a watchfiles watcher,
/media with range support and traversal protection, UI mounts
- backlot/__main__.py: 'python -m backlot open [project]' idempotent
launcher (spawns detached server, opens browser); 'serve' foreground
- verified against real projects: 73 listed, full state for
signal-from-tomorrow, 206 range responses, SSE change push on
filesystem write
Address PR #240 review feedback from @calesthio:
1. dashscope_image: save EVERY returned image URL, not just the first.
The tool advertised multiple_outputs and accepted n>1 but only read
content[0], silently dropping paid outputs. Now collects all image
URLs across choices/content and downloads each to a distinct indexed
path (foo.png -> foo_1.png, foo_2.png, ...). images_generated now
reflects the actual count downloaded.
Per Qwen Cloud docs, a multi-output task is SUCCEEDED if at least one
image is generated; choices with finish_reason != "stop" are skipped
to avoid downloading partial/failed results.
2. Complete idempotency_key_fields so different requests no longer
collide and reuse stale artifacts:
- image: + negative_prompt, seed, prompt_extend, watermark
- tts: + instructions
- asr: + enable_words, language_hints
Adds 19 regression tests (114 total, all pass, no API keys needed):
- TestDashscopeImageMultiOutput: URL extraction across choices / within
one choice / failed-choice skipping, path resolution for
single/multi/no-extension, end-to-end multi-image download with a
mocked 3-URL DashScope response verifying all 3 files land on disk,
single-image legacy path behavior
- TestDashscopeIdempotencyKeys: field presence + key-differs-on-value
for every newly added field across all three tools
Per maintainer feedback on PR #227:
- Revert success=not issues back to success=True — tool execution
succeeded even when QA finds issues; verdict lives in status/issues
- Update test to assert the real contract: success=True + status='revise'
+ issues non-empty, matching how compose-director actually gates
- Consistent with visual_qa.py: success=True, verdict in validation_passed
Addresses review feedback on export_bundle:
- If subtitles_path or thumbnail_path is provided but the file is missing, the
tool now fails with an explicit error instead of silently producing a package
without that asset (which could ship an approved deliverable missing part of
its content).
- Default export location now stays inside the project workspace: when the
render lives at projects/<name>/renders/..., the bundle defaults to
projects/<name>/exports/ (alongside artifacts/, assets/, renders/) rather than
a repo-root exports/<name>/. export_dir remains an explicit override.
Tests cover both: missing optional asset errors, and the project-workspace
default path.
The selector path previously hid the custom-workflow feature: video_selector
filtered tools on per-operation readiness (bundled WAN models) and both
selectors only chose ToolStatus.AVAILABLE providers, so comfyui_image/
comfyui_video — DEGRADED when bundled model metadata is missing — were
dropped even when the ComfyUI server was up and the caller supplied a full
workflow_json/workflow_path plus output_node.
- Add a custom-workflow readiness path to both selectors: when a custom
workflow is supplied, eligibility is based on server availability (status
!= UNAVAILABLE) for any provider advertising supports.custom_workflow,
not on bundled-model readiness. A custom workflow also restricts routing
to custom-workflow-capable providers, since the graph JSON is ComfyUI
specific.
- Expose workflow_json, workflow_path, output_node, workflow_name,
workflow_model, and workflow_model_stack in both selector schemas so
agents can discover the feature without bypassing the selectors.
- image_selector only forwards the workflow inputs to providers that
declare them.
- Add contract tests for the new eligibility path and schema exposure.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Removed comfyui_music and its workflow. The ACE-Step model runs in
ComfyUI but the node class names differ across custom node packs
(AceStepModelLoader vs native TextEncodeAceStepAudio, etc.), so a
bundled workflow would break for most users.
Documented the reasoning in the plan doc and listed it as an open
question for future work. Users with ACE-Step working can still use
the workflow_json override on any tool.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Client queries ComfyUI /object_info to discover installed models
(checkpoints, diffusion models, VAE, CLIP, LoRAs)
- Each tool declares its required models and checks them on execute()
- get_status() returns DEGRADED when server is up but models are missing
- Clear error messages tell the user exactly which models to download
- When COMFYUI_SERVER_URL is not set, error message tells the user to
configure it in .env instead of silently failing on localhost:8188
- 8 new tests covering URL config, error messages, and model requirements
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds three new BaseTool providers that delegate GPU work to a running
ComfyUI server via its REST API. This avoids the need to install
PyTorch/diffusers directly, which is critical on hardware where the
ecosystem hasn't caught up (e.g. NVIDIA Blackwell / DGX Spark, aarch64
+ CUDA 13.0).
New files:
- tools/_comfyui/client.py — shared REST client (submit/poll/download)
- tools/_comfyui/workflows/ — 4 bundled workflow templates
- tools/graphics/comfyui_image.py — FLUX 2 Dev NVFP4 text-to-image
- tools/video/comfyui_video.py — WAN 2.2 14B t2v + i2v (4-step LightX2V)
- tools/audio/comfyui_music.py — ACE-Step 3.5B music generation
- tests/contracts/test_comfyui_tools.py — 41 contract tests
- docs/comfyui-adapter-plan.md — design document
Zero changes to existing tools, selectors, registry, or pipelines.
Tools are auto-discovered and selectors pick them up via capability match.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The contract test in test_character_animation_pipeline.py asserts
result.success after the QA run, implying success should reflect
whether QA passed. But CharacterAnimationReviewer always returned
success=True even when issues were found and status was 'revise'.
This creates a silent failure path: callers that gate on result.success
(the standard ToolResult contract) would treat a broken rig as a clean
pass without ever reading report['status'].
Fix: return success=not issues so result.success is False when QA finds
problems, consistent with the test contract and with how execution errors
are already handled (success=False on exception paths).
Also adds test_character_reviewer_success_false_when_qa_finds_issues to
explicitly assert this contract — a broken rig with missing joints must
produce status='revise', non-empty issues[], and success=False.
Every pipeline ends in a publish stage that produces a publish_log artifact,
but tools/publishers/ shipped empty (only __init__.py) — no tool backed the
PUBLISH tier, so the mechanical packaging (copy the render, write metadata
files, lay out the export directory, emit a schema-valid publish_log) had to be
hand-rolled by the agent every run.
Add a local, offline export bundler:
- capability 'publish', provider 'local', runtime 'local', deterministic, no cost
- takes the final video_path plus the SEO metadata the publish-director prepares
(title, description, tags, hashtags, chapters, optional subtitles/thumbnail)
- writes exports/<project>/{video,metadata,thumbnails}/ matching the
publish-director skill's documented layout
- returns a schema-valid publish_log (status: 'exported') in data, validated
against schemas/artifacts/publish_log before returning so a bad entry fails
here rather than at checkpoint
It does not upload — a networked publisher (e.g. YouTube) can be added later as
a separate provider under the same 'publish' capability.
Tests cover the contract, registry discovery, the export layout, chapter-time
formatting, the schema-valid publish_log, and the missing-video error path.
AGENT_GUIDE.md requires the music decision to be made at the proposal stage,
but the only check for the user's music_library/ folder lived in the
asset-director skills, which run later. A user could approve a creative
direction without ever being told a free, intentional music option was sitting
on disk (issue #168).
music_library/ was already referenced as a source_tool in asset artifacts but
had no backing tool. Add a small read-only tool that scans the library folder
(default <project root>/music_library, override via MUSIC_LIBRARY_DIR or a
library_dir input) and lists the audio tracks it finds, with best-effort
durations via ffprobe when present.
Because it inherits BaseTool, the registry auto-discovers it and it appears in
the preflight provider menu alongside music_gen and the stock music sources:
- AVAILABLE when the folder holds at least one audio track
- UNAVAILABLE otherwise, with install_instructions telling the user how to add
tracks
So the user sees their music options before approving creative direction, with
no orchestration code changes. Read-only: no side effects, no cost.
Closes#168
The compose target resolution was resolved from `profile` (and documented as
overridable via edit_decisions.metadata.compose_target) but the per-segment
scale/pad filter hardcoded 1920x1080. As a result, vertical profiles such as
`tiktok` / `youtube_shorts` / `instagram_reels` silently produced landscape
1920x1080 output instead of 1080x1920 — a silent dimension bug, no error raised.
- Use the resolved target width/height in the per-segment scale/pad filter.
- Implement the previously-stubbed `metadata.compose_target` extension point:
{"width", "height", "fit"} where fit="pad" (letterbox, default, unchanged
behavior) or fit="cover" (scale-to-fill + centre-crop, ideal for vertical).
- Default with no profile/target stays 1920x1080 (backward compatible).
Adds tests/tools/test_video_compose_vertical.py covering default landscape,
profile=tiktok vertical, and compose_target cover override.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
