gewuyou/GFramework
1
0
Fork 0
mirror of https://github.com/GeWuYou/GFramework.git synced 2026-05-07 00:39:00 +08:00
Code Issues Packages Projects Releases Wiki Activity
GFramework/ai-plan
History
gewuyou fdf7382717 fix(sourcegenerators): 收口PR审查遗留问题 
- 修复 PR review 指出的 XML 文档位置、快照路径防御与换行归一化细节

- 更新 monster schema snapshot 场景，覆盖 dependentRequired、dependentSchemas、allOf 与 if/then/else 约束文档

- 补充 SchemaConfigGenerator 条件与组合校验 helper 的 XML 文档，并同步 ai-plan 恢复点与验证记录
2026-04-23 17:38:35 +08:00
..
public
fix(sourcegenerators): 收口PR审查遗留问题
2026-04-23 17:38:35 +08:00
README.md
docs(ai-plan): 收口主题内归档与恢复入口
2026-04-19 15:11:50 +08:00

README.md

AI Plan

ai-plan/ stores AI task recovery artifacts for this repository, but each subdirectory serves a different sharing and bootstrapping purpose.

Directory Semantics

  • public/README.md
    • Shared startup index for boot.
    • Maps worktrees or branches to active topics and points at the primary tracking/trace entry paths.
    • Must list only active topics.
  • public/<topic>/todos/
    • Repository-safe recovery documents for one active topic.
    • Use these for durable task state that another contributor or worktree may need to resume safely.
  • public/<topic>/traces/
    • Repository-safe execution traces for one active topic.
    • Record decisions, validation milestones, and the immediate next step.
  • public/<topic>/archive/
    • Stage-level archive for completed artifacts that still belong to an active topic.
    • Use this when a topic remains active, but some prior phase no longer belongs in the default boot path.
    • Prefer archive/todos/ and archive/traces/ when archiving content cut from the active recovery entry files.
  • public/archive/<topic>/
    • Completed-topic archive.
    • Move the entire topic directory here when that work direction is fully complete.
  • private/
    • Worktree-private recovery space.
    • Use this for temporary notes, local scratch recovery points, or state that only matters in the current worktree.
    • Keep this directory untracked.

Workflow Rules

  • boot must read public/README.md first, then the mapped active topic directories, and only then fall back to scanning active topics directly.
  • If no mapping exists for the current worktree or branch, scan public/<topic>/ and ignore public/archive/ unless the user explicitly asks for historical context.
  • When a worktree changes its active topic set, update public/README.md in the same change.
  • When a stage is complete, move the finished artifacts into public/<topic>/archive/.
  • Keep the active todos/ and traces/ entry files concise. Once a stage is complete and validated, move its detailed history into public/<topic>/archive/ and leave only the current recovery point, active facts, active risks, immediate next step, and archive pointers in the default boot path.
  • When a topic is complete, move the whole topic directory into public/archive/<topic>/ and remove it from the shared startup index.

Content Rules

  • Never write secrets, tokens, credentials, private keys, hostnames, IP addresses, proprietary URLs, or other sensitive data.
  • Never write absolute file-system paths, home-directory paths, or machine usernames.
  • Use repository-relative paths, branch names, PR numbers, recovery-point IDs, and stable document identifiers instead.
  • Keep committed public/** content concise, handoff-safe, and understandable without machine-local context.

Naming

  • Topic directories should be named by capability or work direction, for example:
    • public/ai-plan-governance/
    • public/cqrs-rewrite/
  • Worktree-private files should live under a folder named for the current branch or worktree, for example:
    • private/feat-cqrs-optimization/
    • private/gframework-cqrs/