Skip to contentllms.txt

auto_invocable

Whether the provider supports automatic model-driven skill invocation — the model selects and activates a skill without explicit user syntax. Implementations vary: description-field semantic matching, explicit keyword arrays, and boolean flags are all equivalent mechanisms for the same concept. Note: a description field paired with an opt-out disable_model_invocation flag is functionally equivalent to an explicit auto_invocable boolean — both express that the skill participates in automatic activation unless suppressed.

Type: bool Content type: skills

ProviderSupportedMechanism
amp✓Model auto-selects skill by matching user request against frontmatter name and description fields; no explicit opt-out mechanism is documented.
cline✓Cline auto-invokes skills by matching user message against the description frontmatter field via the use_skill tool; no explicit @-mention syntax exists.
devin✓Cascade auto-invokes a skill when the user's request semantically matches its description frontmatter field.
kiro✓Keywords array in POWER.md frontmatter (key: keywords); Kiro activates the power when any listed keyword is mentioned in conversation.
pi✓Skill name and description are injected into the system prompt as XML; the agent loads matching SKILL.md bodies on demand. Skills with disable-model-invocation: true are excluded from system prompt injection.
zed✓By default the agent picks up skills autonomously: it sees a catalog of every installed skill (name + description) in its system prompt and calls the skill tool when a task matches a skill's description. Semantic description matching, opt-out via disable-model-invocation — functionally the canonical auto_invocable capability.
claude-code✗not documented
codex✗not documented
copilot-cli✗not documented
crush✗not documented
cursor✗not documented
factory-droid✗not documented
gemini-cli✗not documented
opencode✗not documented
roo-code✗not documented