analyze-context

分类: 工具与效率 | 上传者: medhatgalal | 下载: 0 | 版本: v1.0（最新）

[Medhat] Iterative Analysis Workflow for multi-file processing

更新日志: Source: GitHub https://github.com/medhatgalal/Core-Prompts

目录结构

当前层级: .codex/skills/analyze-context/

📄 SKILL.md 2.7 KB

SKILL.md

---
name: "analyze-context"
description: "[Medhat] Iterative Analysis Workflow for multi-file processing"
---
# Iterative Analysis Workflow

## Safety Check
Before proceeding, you MUST check the current branch.
```bash
git status --short --branch
```
If you are on `main` or `master`, STOP. Request `start <slug>`.

This steering file defines a standard workflow for any multi-step analysis task that requires:
- Processing multiple files/items
- Maintaining context across memory compactions
- Accumulating insights iteratively

## Required Memory Files

Before starting any iterative analysis, create these three **Canonical Files** in `.analyze-context-memory/` at the project root.

**CRITICAL RULE: ONE SET ONLY.**
Do NOT create versioned files like `task-v1-context.md` or `task-v2-context.md`.
Maintain **one** active set of files for the entire duration of the initiative. Update them in place.

### 1. `<task>-context.md`
Contains:
- Goal of the analysis (what you're trying to achieve)
- Success criteria (what "done" looks like)
- Scope (what's included/excluded)
- Key definitions or extraction criteria
- Any constraints or rules

### 2. `<task>-todo.md`
Contains:
- Full list of items to process (files, transcripts, etc.)
- Checkbox for each item: `- [ ]` pending, `- [x]` complete
- Notes on what was found per item (brief)
- Current status and next item to process

### 3. `<task>-insights.md`
Contains:
- Accumulated findings, updated after each item
- Patterns emerging across items
- Key quotes, phrases, or data points extracted
- Summary sections that grow as analysis progresses

## Workflow Rules

### Maintenance & Anti-Sprawl
1.  **Single Source of Truth:** Never fork memory files. If the scope changes, update `<task>-context.md`.
2.  **Consolidation:** If you find scattered memory files, merge them back into the canonical set immediately.
3.  **Archival:** Only move files to `docs/archive/` when the *entire initiative* is closed and a new one begins.

### Before Starting
1. Check if `<task>-context.md` already exists. Use it if so.
2. If not, create all three memory files.

### While Working
1. Process one item at a time.
2. After each item:
   - Update insights with new findings.
   - Check off the item in todo.
   - Add brief notes to todo entry.
3. **Critical:** Update files BEFORE memory compaction risk.

### After Memory Compaction
1. **First action:** Read context file to restore goal.
2. **Second action:** Read todo file to find current position.
3. **Third action:** Read insights to restore accumulated knowledge.
4. Continue from where you left off.

### Completion
1. All items checked in todo.
2. Insights contains complete analysis.
3. Summary section in insights with key findings.

登录后下载/点赞/收藏 ❤ 2 | ★ 0

analyze-context

目录结构

SKILL.md

举报内容

提示