manage skills
The workspace’s skills, its saved methods: list, get, create, update, delete, and disable or enable one for this user alone. list returns each skill’s name, purpose and how many colleagues use it, plus the ones switched off; get returns one in full, or several with ids (a skill and its parts in one call). update replaces the whole body, and update and delete reach everyone in the workspace. Model names, prices and tool lists go stale inside a skill. Eversince’s own skills can only be switched off. Free.
Authorizations
An API key (es_live_…) from Settings, under API keys, or an OAuth access token.
Headers
Retry a paid dispatch safely: the same key returns the job already dispatched instead of charging again.
The workspace this one call acts in, when it is not the key's own.
Body
create, list, get, update, delete, disable, enable Skill id from list, required for get, update, delete, disable and enable. A skill's parts switch with it.
For get: several skill ids in one call, returned in the order given; a miss is named under missing. An answer stops at about 100,000 characters and names the rest in next_ids.
1 - 20 elementsFor get: the most characters this answer may hold, for a host that cuts large tool results. One skill is returned up to there, cut at a paragraph, with next_cursor for the rest; several are returned whole up to there.
2000 <= x <= 100000For get with one id: continue a body from the next_cursor a previous answer returned.
x >= 0The kind of job, in a few plain words, distinct from every name in list.
The whole body; update replaces it. Line one is the description agents choose by: one plain sentence, 140 characters or fewer, not a heading. The method follows, up to 25,000 tokens (about 100,000 characters), the most a host hands an agent whole.
Optional: one sentence shown to the user beside this action.
600
