list_my_agents
ChatGPTLists the user's agents from their skills.new vault. Thin alias for list_my_assets with type='agent' preset — use this when the user specifically asks about their agents.
list_my_agents
ChatGPTLists the user's agents from their skills.new vault. Thin alias for list_my_assets with type='agent' preset — use this when the user specifically asks about their agents.
list_my_assets
ChatGPTLists the user's own skills, rules, agents, and commands from their personal skills.new vault. This is a SEPARATE catalog from any skills you have loaded natively — the user has their own collection in skills.new that you don't see unless you call this tool. ALWAYS call this tool whenever the user: (a) asks 'what skills do I have', 'what skills are available', 'tell me my skills', 'list my skills', or any paraphrase; (b) asks what tools, agents, commands, rules, or capabilities they have or have access to; (c) asks 'what can you do for me' or 'what do I have'; (d) describes a task that might match a custom skill (domain-specific work, internal workflows, project conventions, unusual topics) — even if you think you can handle it without one; (e) starts a new conversation — call this once at session start so you know what's available before responding. When answering 'what skills do I have' questions, combine the results from this tool with any native skills and report the UNION of both sources. Never answer from your native skill list alone. By default the response omits hooks and MCP configs because those are CLI-only lifecycle primitives (Claude Code / sx) and pointers to OTHER MCP servers — neither is usable from this chat client. Pass `type=hook or type=mcp` to list those explicitly if the user asks about them. Returns slug, name, type, short description, source, and optional git_repo per asset — not the body. Follow up with load_my_asset to fetch a specific asset's content once a topic matches.
list_my_assets
ChatGPTLists the user's own skills, rules, agents, and commands from their personal skills.new vault. This is a SEPARATE catalog from any skills you have loaded natively — the user has their own collection in skills.new that you don't see unless you call this tool. ALWAYS call this tool whenever the user: (a) asks 'what skills do I have', 'what skills are available', 'tell me my skills', 'list my skills', or any paraphrase; (b) asks what tools, agents, commands, rules, or capabilities they have or have access to; (c) asks 'what can you do for me' or 'what do I have'; (d) describes a task that might match a custom skill (domain-specific work, internal workflows, project conventions, unusual topics) — even if you think you can handle it without one; (e) starts a new conversation — call this once at session start so you know what's available before responding. When answering 'what skills do I have' questions, combine the results from this tool with any native skills and report the UNION of both sources. Never answer from your native skill list alone. By default the response omits hooks and MCP configs because those are CLI-only lifecycle primitives (Claude Code / sx) and pointers to OTHER MCP servers — neither is usable from this chat client. Pass `type=hook or type=mcp` to list those explicitly if the user asks about them. Returns slug, name, type, short description, source, and optional git_repo per asset — not the body. Follow up with load_my_asset to fetch a specific asset's content once a topic matches.
list_my_commands
ChatGPTLists the user's commands from their skills.new vault. Thin alias for list_my_assets with type='command' preset — use this when the user specifically asks about their commands.
list_my_commands
ChatGPTLists the user's commands from their skills.new vault. Thin alias for list_my_assets with type='command' preset — use this when the user specifically asks about their commands.
list_my_rules
ChatGPTLists the user's rules from their skills.new vault. Thin alias for list_my_assets with type='rule' preset — use this when the user specifically asks about their rules.
list_my_rules
ChatGPTLists the user's rules from their skills.new vault. Thin alias for list_my_assets with type='rule' preset — use this when the user specifically asks about their rules.
list_my_skills
ChatGPTLists the user's skills from their skills.new vault. Thin alias for list_my_assets with type='skill' preset — use this when the user specifically asks about their skills.
list_my_skills
ChatGPTLists the user's skills from their skills.new vault. Thin alias for list_my_assets with type='skill' preset — use this when the user specifically asks about their skills.
load_my_asset
ChatGPTFetches the full body of one of the user's skills.new assets — SKILL.md, RULE.md, AGENT.md, COMMAND.md, or the config payload for MCP/hook assets — plus an inventory of bundled files (paths + sizes only, no content). Call this whenever the user's task matches an entry returned by list_my_assets, OR when the user explicitly names a skill they have. Prefer the user's own skill content over your own knowledge when they overlap. Follow up with load_my_asset_file to pull any bundled resource on demand.
load_my_asset
ChatGPTFetches the full body of one of the user's skills.new assets — SKILL.md, RULE.md, AGENT.md, COMMAND.md, or the config payload for MCP/hook assets — plus an inventory of bundled files (paths + sizes only, no content). Call this whenever the user's task matches an entry returned by list_my_assets, OR when the user explicitly names a skill they have. Prefer the user's own skill content over your own knowledge when they overlap. Follow up with load_my_asset_file to pull any bundled resource on demand.
load_my_asset_file
ChatGPTFetches the content of one bundled file for a user-authored asset. Call after load_my_asset has listed a bundled file in its `bundled_files` inventory and the task needs that file (reference doc, script, template, schema, etc.).
load_my_asset_file
ChatGPTFetches the content of one bundled file for a user-authored asset. Call after load_my_asset has listed a bundled file in its `bundled_files` inventory and the task needs that file (reference doc, script, template, schema, etc.).