refactor: split skill providers

examples/acp-agent/tests/snapshots/skill-load/stdout.golden.jsonl

@@ -2,7 +2,7 @@
 {"jsonrpc":"2.0","id":2,"result":{"sessionId":"{{sessionId}}"}}
 {"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"agent_thought_chunk","content":{"type":"text","text":"Load the requested skill."}}}}
 {"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call","toolCallId":"call_skill_load","title":"Load skill dsh-skill-creator","kind":"read","status":"in_progress","rawInput":"dsh-skill-creator"}}}
-{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call_update","toolCallId":"call_skill_load","status":"completed","content":[{"type":"content","content":{"type":"text","text":"<skill_content name=\"dsh-skill-creator\">\n# Skill: dsh-skill-creator\n\nUse this skill to write focused DeepSeek Harness skills.\n\nA skill is a directory `<name>/SKILL.md` or a flat `<name>.md` file with YAML frontmatter.\nFrontmatter must include kebab-case `name` and a concise `description` that tells the model when to load it.\nUse optional `whenToUse` for extra routing signal and `disableModelInvocation: true` for user-only skills.\nKeep the body procedural, evidence-oriented, and scoped to the workflow the skill owns.\n\nBase directory for this skill: {{cwd}}/.dsh/skills/.system/dsh-skill-creator\nResolve relative files mentioned by this skill against the base directory before using them.\n</skill_content>"}}]}}}
+{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call_update","toolCallId":"call_skill_load","status":"completed","content":[{"type":"content","content":{"type":"text","text":"<skill_content name=\"dsh-skill-creator\">\n# Skill: dsh-skill-creator\n\nUse this skill to write focused DeepSeek Harness skills.\n\nA skill is a directory `<name>/SKILL.md` or a flat `<name>.md` file with YAML frontmatter.\nFrontmatter must include kebab-case `name` and a concise `description` that tells the model when to load it.\nUse optional `whenToUse` for extra routing signal and `disableModelInvocation: true` for user-only skills.\nKeep the body procedural, evidence-oriented, and scoped to the workflow the skill owns.\n\nBase directory for this skill: {{cwd}}/.dsh/skills/dsh-skill-creator\nResolve relative files mentioned by this skill against the base directory before using them.\n</skill_content>"}}]}}}
 {"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"agent_thought_chunk","content":{"type":"text","text":"The skill is loaded."}}}}
 {"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"agent_message_chunk","content":{"type":"text","text":"DONE"}}}}
 {"jsonrpc":"2.0","id":3,"result":{"stopReason":"end_turn"}}

examples/acp-agent/tests/snapshots/skill-load/workspace/.dsh/skills/dsh-skill-creator/SKILL.md

@@ -0,0 +1,12 @@
+---
+name: dsh-skill-creator
+description: Create or update DeepSeek Harness SKILL.md instructions.
+whenToUse: Use when writing reusable agent instructions for DeepSeek Harness or adding system/project/user skills.
+---
+
+Use this skill to write focused DeepSeek Harness skills.
+
+A skill is a directory `<name>/SKILL.md` or a flat `<name>.md` file with YAML frontmatter.
+Frontmatter must include kebab-case `name` and a concise `description` that tells the model when to load it.
+Use optional `whenToUse` for extra routing signal and `disableModelInvocation: true` for user-only skills.
+Keep the body procedural, evidence-oriented, and scoped to the workflow the skill owns.