Skip to main content
GET
Get the body markdown of a skill for prompt injection / load_skill

Authorizations

Authorization
string
header
required

Enter your API key (e.g. dai_sk_test_key_001)

Headers

Authorization
string

Path Parameters

skill_name
string
required

Query Parameters

version
integer | null

Version number (default: the version the menu offers)

agent_name
string | null

Resolve the exact version this agent was offered (agent-scope Use pins).

max_chars
integer | null

Trim the body to this many characters (progressive-disclosure body guardrail — the SDK's load_skill passes its per-body limit).

Required range: x >= 256
routing_id
string | null

The routing decision (routing_id from POST /skills/route or GET /skills/menu) this load belongs to; the SDK's load_skill sends it. With it, a skill you can reach only through the public registry (not owned by your organization, installed in it, or assigned to the agent) is served only if that decision offered it, and answers 404 otherwise. Skills your organization owns, has installed or has assigned are served either way. When agent_name is omitted, or names an agent that does not hold the skill, the load is resolved for the agent the decision was made for, so that agent's pinned version is the one served. Send it empty from a run whose routing call returned no decision. Omit it outside an agent run (export, skills pull).

Maximum string length: 64

Cookies

decimal_session
string | null

Response

Successful Response