- Home
- Skills
- Workflow Automation
- Team Habits for Coding Agents
Team Habits for Coding Agents
AI coding agents fail in the same ways human teams always have.
Free
Works with the AI Agents you already use
Team Habits for Coding Agents
Example session with this skill installed
Use the glossary-and-actors skill. I'm Sam and I approve all work in this repo. The README calls them tasks, but I sometimes say todos.
- Read your context and instructions
- Compiled the team habits
Created docs/glossary.md with a People and roles table (Sam: requests and approves the agent's work, write "Sam" in records, not "the user"; User: a person who uses the app) and a Terms table (Task, aliases: todo, item). Linked it from README.md.
Connects securely to your tools. The creator never sees your data.
About this skill
AI coding agents fail in the same ways human teams always have. A suggestion quietly becomes a decision. Notes pile up until nobody knows which line is still true. One word means two different things. Work starts on whatever looks easiest.
Software teams found fixes for these long ago: decision records, current-state documents, a shared glossary, dependency-based ordering, independent review. This pack teaches your agent those habits, one skill per failure, and tells it where to write things down so the next session can pick up where this one stopped.
Skills
| Skill | Failure it prevents | What it leaves in your repo |
|---|---|---|
| start-here | The agent never loads the other skills | A short table in CLAUDE.md / AGENTS.md that maps situations to skills |
| requirements-vs-means | A proposed means is treated as a requirement | Decisions labelled Confirmed, Provisional or Open |
| current-state-docs | Appended corrections bury what is true now | Rewritten documents, history moved to reports, docs/known-gaps.md |
| glossary-and-actors | One name for two things, or two names for one | docs/glossary.md, including the people involved |
| ordering-by-dependency | Work starts on the easiest or lowest-numbered task | A dependency-ordered task list with impact-based priorities |
| plan-refutation | The agent approves its own plan | One recorded refutation pass by a fresh reviewer |
| record-placement | Decisions are written where the next session never looks | Each record in the place that will be read when it is needed |
| verify-in-operation | A mechanism is called done after it worked once, in one place | A per-target table of what was checked, how, and what was not checked |
Start with start-here
Free skills are not announced to your agent automatically. Run start-here once per project: it writes a small table into CLAUDE.md and AGENTS.md, which agents read at the start of every session. From then on the agent loads the right skill when the situation comes up, without you naming it.
Source and feedback
The skills are open source under the MIT License at github.com/AllegroMoltoV/team-habits-for-coding-agents. Questions and bug reports are welcome in its Issues. Release notes: github.com/AllegroMoltoV/team-habits-for-coding-agents/releases They are adapted from the author's earlier rule set for Claude Code, rewritten as on-demand, agent-neutral skills.
Skills are loaded on demand, so they only apply in situations listed by start-here or when you name them. Nothing is enforced by hooks. Beads (bd) is optional; without it the skills use Markdown files under docs/.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 15 days ago
- Free with an Agensi account
Requires
Permissions
File Scopes
Writes only Markdown files in your repository. No network access, no scripts.