一键导入
skill-cslib-research
Research CSLib formalization patterns and Mathlib API for CSLib contributions. Invoke for cslib research tasks.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Research CSLib formalization patterns and Mathlib API for CSLib contributions. Invoke for cslib research tasks.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Orchestrate multi-agent implementation with parallel phase execution. Spawns teammates for independent phases and coordinates dependent phases. Includes debugger teammate for error recovery.
Orchestrate multi-agent planning with parallel plan generation. Spawns 2-3 teammates for diverse planning approaches and synthesizes into final plan with trade-off analysis.
Orchestrate multi-agent research with wave-based parallel execution. Spawns 2-4 teammates for diverse investigation angles and synthesizes findings.
Full structural hard-mode orchestration state machine with per-phase dispatch (H1), adversarial verification (H4), convergence policing (H6), territory contracts (H7), and churn detection (H5). Invoke for /orchestrate --hard.
Autonomous state machine that drives a task through its full lifecycle (research -> plan -> implement -> complete) without user confirmation between phases. Invoke for /orchestrate command.
Execute hard-mode implementation with anti-analysis contracts, per-phase dispatch, and territory-aware execution. Invoke for --hard implementation tasks.
| name | skill-cslib-research |
| description | Research CSLib formalization patterns and Mathlib API for CSLib contributions. Invoke for cslib research tasks. |
| allowed-tools | Agent, Bash, Edit, Read, Write |
Thin wrapper that delegates CSLib research to cslib-research-agent subagent.
This skill activates when:
Validate task_number exists and task_type is "cslib".
Update status to "researching" BEFORE invoking subagent.
Domain-specific context for the cslib-research-agent:
.claude/extensions/cslib/context/{
"session_id": "sess_{timestamp}_{random}",
"delegation_depth": 1,
"delegation_path": ["orchestrator", "research", "skill-cslib-research"],
"timeout": 3600,
"task_context": {
"task_number": N,
"task_name": "{project_name}",
"description": "{description}",
"task_type": "cslib"
},
"focus_prompt": "{optional focus}",
"metadata_file_path": "specs/{NNN}_{SLUG}/.return-meta.json"
}
Retrieve relevant memories and literature briefing to inject into the delegation context.
Skip memory if: clean_flag is true (from --clean command flag).
# Check clean_flag
if [ "$clean_flag" != "true" ]; then
memory_context=$(bash .claude/scripts/memory-retrieve.sh "$description" "$task_type" "$focus_prompt" 2>/dev/null) || memory_context=""
fi
# memory_context will be empty string if:
# - clean_flag is true (skipped)
# - memory-index.json missing or empty
# - no keywords matched any entries
# - script exited with error
# Literature briefing injection (independent of clean_flag)
lit_context=""
if [ "$lit_flag" = "true" ]; then
lit_context=$(bash .claude/scripts/literature-briefing.sh 2>/dev/null) || lit_context=""
fi
# lit_context will be empty string if:
# - lit_flag is not "true" (skipped)
# - specs/literature/ sub-index is empty or missing
# - script exited with error
Note: lit_flag is independent of clean_flag. Using --clean --lit suppresses memory retrieval but still injects literature briefing. Literature briefing is gated solely on lit_flag == "true".
Use Agent tool with subagent_type: "cslib-research-agent".
Include memory_context and lit_context in the prompt if non-empty:
memory_context is non-empty, include it as a <memory-context> block after the delegation context.lit_context is non-empty, include it as a <literature-briefing> block after the memory context.CRITICAL: If you performed the work above WITHOUT using the Agent tool (i.e., you read files,
wrote artifacts, or updated metadata directly instead of spawning a subagent), you MUST write a
.return-meta.json file now before proceeding to postflight. Use the schema from
return-metadata-file.md with the appropriate status value for this operation.
If you DID use the Agent tool, skip this stage -- the subagent already wrote the metadata.
The following stages MUST execute after work is complete, whether the work was done by a subagent or inline (Stage 4b). Do NOT skip these stages for any reason.
Read the metadata file from specs/{N}_{SLUG}/.return-meta.json.
Update state.json and TODO.md based on result.
Add research artifact to state.json. Update TODO.md per @.claude/context/patterns/artifact-linking-todo.md with field_name=**Research**, next_field=**Plan**.
Commit changes with session ID.
Brief text summary (NOT JSON).