find-task-skills
Find Task SKILLs
Purpose
You are a teammate in an AITC workflow execution. Task SKILLs — concrete operational knowledge discovered during this work session — exist as markdown files in skills/aitc-task-<batch>/. Your job is to find and load the ones relevant to your task before you begin work.
This skill replaces the Lead manually listing task SKILL references in your prompt. You discover them yourself.
When to Use
Invoke this skill at these points during your work:
| When | Why |
|---|---|
| Before starting any task phase | Load known pitfalls and procedures before you hit them |
| Encountering an error or unexpected behavior | A previous teammate may have already solved this |
| About to use a tool or system you haven't worked with | Someone may have documented quirks |
| After task-skills-creator creates a new task SKILL | Verify it appears in the directory listing |
Procedure
Step 1 — Identify the Active Task SKILL Directory
Your prompt should specify the batch name. If it doesn't, ask the Lead: "Which task SKILL directory should I search?"
The directory is at skills/aitc-task-<batch>/.
Note: skills/ contains only ONE active aitc-task-xxx directory at a time — the one matching the batch name in your prompt. Inactive task SKILL directories from past sessions are in archived/. Do NOT load task SKILLs from archived/; they belong to different work sessions and their information may be stale.
Step 2 — List Available Task SKILLs
Run the listing script to get a structured overview of all task SKILLs in the active directory:
bash plugins/aitc-workflow/skills/find-task-skills/list-task-skills.sh <batch-name>
Or for machine-parseable output:
bash plugins/aitc-workflow/skills/find-task-skills/list-task-skills.sh --json <batch-name>
The script handles these cases automatically:
| Scenario | Script Behavior |
|---|---|
One aitc-task-xxx directory matches the batch name |
Lists all SKILL.md files with frontmatter extracted |
Multiple aitc-task-xxx directories |
Reports ALL directories with creation/modified times and skill counts. Exits with code 1. Asks the agent to identify the active one. |
No aitc-task-xxx directory |
Reports empty, exits 0 |
| Batch name given, directory not found | Reports not found, exits 1 |
If the script reports multiple directories, this is an anomaly — skills/ should only contain one active task SKILL directory. Ask the Lead: "Multiple aitc-task directories found. Which is active? Should the others be moved to archived/?"
If the directory is empty (no .md files), report: "No task SKILLs available yet." and proceed with your task using only the knowledge in your prompt.
Step 3 — Judge Relevance
The script already extracts name, description, task-type, status, supplements/instance-of from each SKILL's frontmatter. Review the output:
- Skip any file with
status: supersededorstatus: merged— these have already been absorbed. - For each active task SKILL, judge: does this relate to my assigned task? Read the description. If the task SKILL covers a tool, system, or operation you'll be working with, it's relevant. If it's about an unrelated domain, skip it.
Step 4 — Load Relevant Task SKILLs
For each task SKILL you judged relevant, read the ENTIRE file. Pay special attention to:
## Prerequisites— what must be true before you can use this knowledge## Procedure— step-by-step instructions; values are CONCRETE and REAL, not placeholders## Parameterization(instance type) — actual parameters for this session## Corrections / Additions(supplement type) — what the original skill got wrong
Follow the procedure exactly. Do not second-guess concrete values — they were placed there by someone who discovered them through direct experience.
Step 5 — Apply Self-Maintenance Rule
Each task SKILL ends with a ## Self-Maintenance Rule. If you find any information in a loaded task SKILL that doesn't match reality:
- Edit the file directly — correct the outdated section
- Append to
## Discoveries— record what you found and what you changed - Report in your completion message — under
## Discoveries > Self-Maintenance
Do not silently work around incorrect information. Fix it.
Step 6 — Report
In your first completion message to the Lead, include:
Task SKILLs loaded:
- <name>: <one-line summary> (relevant because: <reason>)
- <name>: <one-line summary> (relevant because: <reason>)
- None (if no relevant task SKILLs found)
This lets the Lead verify you found and loaded the right knowledge.
More from satone7/skills
writing-plans-plus
Enhanced planning with structured, machine-readable task format based on writing-plans. Use this skill whenever you need to create implementation plans with explicit completion tracking, cross-validation issue tracking, or machine-readable JSON output. Always use this when planning software development tasks, especially when tasks need to be executed by other agents or reviewed later.
10find-next-task
Find the next executable task from writing-plans-plus JSON plan files. Use this skill whenever you need to locate the next task to work on, check task status, or understand what to do next in a project with structured plans. Always use this when the user asks "what's next", "find the next task", "show pending tasks", or wants to continue working on a plan.
5aitc-workflow
>
5guardian
|
4executing-single-task
执行 writing-plans-plus 计划中的“恰好一个任务”:输入为 find-next-task 的完整输出 JSON。严格按四阶段流程(实现→规范审查→质量审查→最终验证),基于证据更新 plan JSON 的 passes(成功时删除 issue 字段;失败时写入/更新非空 issue[]),并创建且仅创建一个 git commit。凡用户要求“执行下一个任务/只跑一个任务/把 passes 写回/自动执行单任务”都应优先使用本技能。
4github-pr-fixer
>
3