- Home
- Skills
- Research & Analysis
- PhytoSkill-Spark — 5-Skill Research Demo
PhytoSkill-Spark — 5-Skill Research Demo
It does **not** inspect arbitrary real images, diagnose plant disease, identify pathogens, or produce validated agricultural recommendations.
Free
Works with the AI Agents you already use
PhytoSkill-Spark — 5-Skill Research Demo
Example session with this skill installed
User task: "Run the built-in synthetic Huangqi leaf observation demo; do not analyze a real image."
The agent creates request.json with this exact input:
{
"case_id": "demo-huangqi-001",
"species": "黄芪",
"image_path": "fixture://huangqi-leaf-01"
}
Then runs the plant-vision example:
python run.py --input request.json
- Read your context and instructions
- Compiled the phytoskill-spark
{
"status": "success",
"case_id": "demo-huangqi-001",
"species": "黄芪",
"provenance": {
"data_origin": "synthetic_fixture",
"skill": "plant_vision",
"version": "0.2.0",
"model_called": false
},
"observations": [
{
"observation_id": "obs-yellow-01",
"phenotype": "leaf_yellowing",
"region": {
"label": "leaf edge",
"bbox_normalized": [
0.2,
0.15,
0.8,
0.85
]
},
"affected_area_fraction": 0.21,
"area_basis": "visible_leaf_area",
"model_score": 0.92
}
],
"limitations": [
"本结果来自合成fixture,未读取图片或执行模型推理。",
"表型和模型分数不能用于确诊病原;面积比例不是药效下降比例。"
]
}
The result is deterministic and labeled synthetic_fixture. An unmatched input results in an error; it does not trigger a real vision model.
Connects securely to your tools. The creator never sees your data.
About this skill
Demo / Research Prototype. Not a production agricultural diagnostic system.
PhytoSkill-Spark is a free five-skill learning and engineering demonstration using an explicitly synthetic medicinal-plant fixture (Huangqi / 黄芪). It shows how a set of composable agent skills can pass structured outputs, preserve evidence IDs, and flag unsupported report claims. It does not inspect arbitrary real images, diagnose plant disease, identify pathogens, or produce validated agricultural recommendations.
Included:
plant-vision: returns a fixed synthetic leaf-yellowing observation with a traceable observation ID, region coordinates and a model-score field (not a disease probability).growth-risk: returns synthetic measurements and possible factors with missing context retained; it does not infer soil or nutrient conditions from an image.herbal-knowledge: returns one synthetic evidence entry for interface testing. No corpus, live search or actual literature verification is included.evidence-fusion: links supplied observation, environment and knowledge JSON by IDs and case/species, retains missing sources and labels relationshipsco_occurrence_only. Its input authenticity restriction is now enforced: only the published fixture sources, verbatim, are accepted; any non-fixture or altered source is refused with exit 2 and no output.agentshield-audit: runs deterministic metadata- and ID-consistency checks on caller-provided package manifests or reports. It returnsPASS,INCOMPLETE,INSUFFICIENT, orQUARANTINEas appropriate. It does not verify cryptographic file integrity, validate external scientific truth or enforce runtime permissions.
The original research repository has a separate signed registry and middleware, neither of which is included in this free stand-alone distribution pack. This pack includes no GPU inference, API keys, external models, network calls or third-party corpus. Results should be treated as demonstrations of data contracts and fail-closed checks, not verified scientific findings.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Research & Analysis.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 18 days ago
- Free with an Agensi account