You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
"Look up skill details by capability. Skill names are listed in the system prompt; call this for descriptions, then use_skill to load a body. Directly callable — do not tool_search for this.",
35
+
"Look up skill details by capability, then use_skill to load a body. Directly callable. See Guidelines: Tool choice.",
"Find a skill during prep when attached skills are not enough. Do not search on a tiny one-file fix. Directly callable — do not tool_search for this. Returns name + description; load a body with use_skill.",
42
+
"Prep lookup for a skill when attached skills are not enough; skip tiny one-file fixes. Directly callable; load via use_skill. See Guidelines: Tool choice.",
42
43
inputSchema: SKILL_SEARCH_INPUT_SCHEMA,
43
44
};
44
45
@@ -105,7 +106,7 @@ export function createSkillSearchTool(
105
106
return`No skills matched "${query}". Try different keywords describing the capability.`;
"Load the full instructions for a skill. Names are listed under 'Skills' in the system prompt; call skill_search for descriptions, then this tool with the skill's name to load the body. The returned instructions stay in effect for the rest of the task.",
32
+
"Load the full instructions for a skill; locate details first with skill_search. See Guidelines: Tool choice.",
"Load a skill you already know by name (brief or search). Do not reload skills listed as attached or already in context. The returned instructions stay in effect for the rest of the task.",
39
+
"Load a skill by name; do not reload attached skills or bodies already in context. See Guidelines: Tool choice.",
"Start a worker agent and return IMMEDIATELY with its agent_id — this never blocks on the worker's completion. Pass agent= a director/profile id returned by search_agents, or intent= (one of explore|implement|review|plan|general). The child starts blank. One focused task per worker. success_criteria is required for implement/review (and their default directors). Fire several spawn_agent calls in one turn to start independent lanes in parallel, then reply and end the turn — workers keep running while you are idle. Reports arrive as mailbox mail where mailbox delivery is mounted; where wait_agents is mounted (exec primary), collect with it instead. Do not poll. Excess fan-out is queued rather than refused.",
551
+
"Start a worker agent and return immediately with its agent_id; one focused task per worker. See Guidelines: Orchestration.",
0 commit comments