CD IMA Studio Video Generation
Use for AI video generation via IMA Open API. Supports 4 modes: text-to-video (14 models), image-to-video (14 models), first_last_frame_to_video (10 models), reference_image_to_video (9 models). IMPORTANT — Default model selection rule: always recommend the NEWEST and most POPULAR model, NOT the cheapest. Default text_to_video: Wan 2.6 (wan2.6-t2v, 25pts) — most popular, balanced cost. Alternative: Hailuo 2.3 (MiniMax-Hailuo-2.3, 38pts) for higher quality. Default image_to_video: Wan 2.6 (wan2.6-i2v, 25pts) — most popular for i2v. Default first_last_frame_to_video: Kling O1 (kling-video-o1, 48-120pts). Default reference_image_to_video: Kling O1 (kling-video-o1, 48-120pts). Production models (2026-02-27) — text_to_video (14): Wan 2.6, Hailuo 2.0/2.3, Vidu Q2, SeeDance 1.5 Pro, Sora 2 Pro, Kling O1/2.6, Google Veo 3.1, Pixverse V3.5-V5.5. image_to_video (14): Same as text_to_video except Vidu Q2 Pro. first_last_frame_to_video (10): Hailuo 2.0, Kling O1/2.6, Vidu Q2 Pro, Google Veo 3.1, Pixverse V3.5-V5.5. reference_image_to_video (9): Kling O1, Google Veo 3.1, Vidu Q2, Pixverse (all versions). Poll every 8s. Requires an ima_* API key.
Use for AI video generation via IMA Open API.
As a process D 37/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.
Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".
If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.
How to improve
- Shorten the description to 1024 characters.
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 3
✓ No critical or high findings
Medium and low: 3
-
medium Dangerous commands
cmd-shell-rcINSTALL.md:42Writes to a shell startup fileecho 'export IMA_API_KEY=ima_your_key_here' >> ~/.bashrc
-
medium Dangerous commands
cmd-shell-rcINSTALL.md:335Writes to a shell startup fileecho 'export IMA_API_KEY=ima_your_key_here' >> ~/.bashrc
-
medium Exfiltration
net-redirectable-api-keyscripts/ima_video_create.py:939Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
Files scanned: 9. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1151 chars, limit 1024 - warning
name-formatname should be kebab-case (lowercase letters, digits, hyphens) - warning
body-longSKILL.md body ≈ 11512 tokens (recommended < 5000); move details to references/ - note
frontmatter-keyunknown frontmatter key "keywords"
Process rating: all ten parameters 37/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 28 mutating operations with no state check
- 40Consistency. Frontmatter name (IMA Studio Video Generation) differs from the folder (ima-video-ai)
- 40Execution cost. Instruction body is 11512 tokens: crowds the task out of the window
- 55Failures and branches. 1 branches
- 60Tools and files. Uses tools (web, python, node) that frontmatter does not declare
- 60Steps. 84 steps, 4 vague phrases
- 100Progress reporting. Reports progress
- low The response is described with custom markup (5 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 1150: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -268 emoji in the instructions: noise for the model
- +1No license
- +2Single-language instructions
- +4Structure: 26 headings
- +3Step-by-step instructions: 84 items
- +4Has examples (25 code blocks)
- +3All 2 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 41.