Skip to content

skill_get

Analysis

Read a Workflow Skill

Read Scenario's workflow skills: tested procedures for common jobs, such as letter-perfect text overlays and banners, product shots, sprite animation, orbit views, textures, 3D, storyboards, video ads, captions, music videos, model training, brand kits and team analytics. Call it when planning a multi-step task, before picking models: a skill names the models, parameters and checks that work, and the traps to avoid. With no name it lists every skill with a one-line summary. With a name it returns that skill's SKILL.md. With a path it returns a markdown file the skill links to, resolved like the link itself, so "../scenario/references/shared-assets.md" reaches the other skill. Returns: the file as markdown (response_format "json" wraps it as { name, path, uri, text, files, unserved }, or { skills } for the list). A skill that bundles scripts or data files says so at the top: running them needs a local install (npx skills add scenario-labs/skills --skill <name>). Examples: skill_get() for the list; skill_get(name="scenario") for the connection and core loop; skill_get(name="scenario-text-overlay"). The same files are MCP resources (skill://scenario/<name>/SKILL.md) for hosts that read resources. Don't use it to find a model for a task (recommend, search) or to read a model's parameters (model_schema_get).
read-onlyidempotent

Parameters

NameTypeRequiredDescription
team_idstring—Optional. Does not affect which skill is served.
project_idstring—Optional. Does not affect which skill is served.
response_formatenum(json | markdown)markdownOutput format: 'markdown' (default) returns the file itself; 'json' wraps it with its metadata.
namestring—Skill name, e.g. "scenario-text-overlay". Omit to list every skill with a one-line summary.
pathstring—A markdown file the skill links to, as the link reads (e.g. "references/setup.md" or "../scenario/references/shared-assets.md"). Default: the SKILL.md.