BA alibabacloud-oss-media-process
Process images, audio, and video files stored in Alibaba Cloud OSS. Supports 14+ image operations (resize, crop, rotate, watermark, blur, format conversion, etc.), image-intelligent features via IMM (blind watermark, face/body/car detection, QR recognition, labeling, scoring), and audio/video processing (transcoding, screenshot, animation, sprite sheet, concatenation, metadata extraction, HLS streaming). Results can be returned as signed URL, downloaded locally, or saved as new OSS object. Also supports plain file upload/download. Use when the user needs to process or transform media files in OSS, such as generating thumbnails, transcoding video, extracting audio, adding watermarks, detecting faces, compressing images, or converting formats. Triggers on media processing requests in English or Chinese (resize, crop, thumbnail, transcode, video convert, audio convert, watermark, face detection, 缩略图, 裁剪, 压缩, 转码, 视频转换, 音频处理, 水印, 盲水印, 人脸检测, 截帧, 拼接).
As a process A 88/100 · Runs to the end — weak spots: execution cost, running it twice
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 skill asks for more than the task needs: broad tool access, credential environment variables, binaries. Every extra permission widens the damage from a mistake or a compromise.
Narrow allowed-tools and the variable list to the minimum; replace binaries with readable sources or scripts.
How to improve
- 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 · 4
✓ No critical or high findings
Medium and low: 4
-
medium Dangerous commands
cmd-autorun-instructionreferences/troubleshooting.md:13Instructs the agent to auto-run a script on every session| Running video/audio/IMM ops without `--imm-project` | `task_id: null`, silent failure, or need to retry | Always run `imm_admin.py auto-setup` first to get project name, then include `--imm-project
-
medium Broad scope
meta-broad-allowed-toolsSKILL.md:1Broad tool permissions pre-approved: Bashallowed-tools: Bash Read
-
medium Dangerous commands
cmd-pipe-to-shellSKILL.md:71Downloads and executes remote code from an unrecognised host (pipe to shell) (documentation table row)| 1 | Install Aliyun CLI (>=3.3.3) | `curl -fsSL https://aliyuncli.alicdn.com/setup.sh | bash` |
table -
medium Dangerous commands
cmd-autorun-instructionSKILL.md:158Instructs the agent to auto-run a script on every sessionBefore running video, audio, HLS, or image-intelligent operations, first run `imm_admin.py auto-setup` to ensure the bucket is bound to an IMM project. Pass `--imm-project <project_name>` only for `bl
Files scanned: 20. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 8772 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 88/100
- 30Running it twice. 8 mutating operations with no state check
- 40Execution cost. Instruction body is 8772 tokens: crowds the task out of the window
- 70Inputs and preconditions. Inputs and preconditions are listed
- 85Steps. 65 steps, 1 vague phrases
- 100Tools and files. Tools declared in frontmatter
- 100Result and completion. Output format and completion criterion are stated
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 26 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 15 top-level sections: this looks like several domains in one skill
- low The response is described with custom markup (10 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 958: 120–800 characters recommended
- -34 of 8 scripts are never mentioned in SKILL.md
- +2Single-language instructions
- +4Structure: 40 headings
- +3Step-by-step instructions: 65 items
- +3Output format is stated explicitly
- +4Has examples (13 code blocks)
- +4Reference files are cited in the instructions (7 of 9)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 76.