agent-teams-subagent-startup-convention

Sub-skill of agent-teams: Subagent startup convention (+2).

5 stars

Best use case

agent-teams-subagent-startup-convention is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Sub-skill of agent-teams: Subagent startup convention (+2).

Teams using agent-teams-subagent-startup-convention should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

You only need a quick one-off answer and do not need a reusable workflow.
You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/subagent-startup-convention/SKILL.md --create-dirs "https://raw.githubusercontent.com/vamseeachanta/workspace-hub/main/.agents/skills/_archive/workspace-hub/agent-teams/subagent-startup-convention/SKILL.md"

Manual Installation

Download SKILL.md from GitHub
Place it in .claude/skills/subagent-startup-convention/SKILL.md inside your project
Restart your AI agent — it will auto-discover the skill

How agent-teams-subagent-startup-convention Compares

Feature / Agent	agent-teams-subagent-startup-convention	Standard Approach
Platform Support	Not specified	Limited / Varies
Context Awareness	High	Baseline
Installation Complexity	Unknown	N/A

Frequently Asked Questions

What does this skill do?

Sub-skill of agent-teams: Subagent startup convention (+2).

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# Subagent startup convention (+2)

## Subagent startup convention


Set this as the **first Bash call** in every non-trivial subagent session:

```bash
export CLAUDE_SUBAGENT=1
```


## Hook behaviour by context


| Hook | Main session | Subagent | Reason |
|------|-------------|----------|--------|
| `session-review.sh` | Runs | Runs | Writes signals to `pending-reviews/` |
| `post-task-review.sh` | Runs | Runs | Captures findings to `pending-reviews/` |
| `consume-signals.sh` | Runs | Runs | Processes signals into session-briefing |
| `improve.sh` | Runs | **Skipped** | Narrow context produces low-quality output |
| `query-quota.sh` | Runs | **Skipped** | Quota tracking is main-session only |

The shared `.Codex/state/pending-reviews/` directory is the passive message
bus. Subagents write signals via the lighter hooks; the main session's Stop
drains them all via `consume-signals.sh` → `improve.sh` at session end.


## Result


Main orchestrator stays lean and interactive. User can inject messages and
redirect subagents mid-flight. Learnings accumulate in the background without
adding any overhead to the orchestrator between Task completions.

Related Skills

subagent-write-verification

from vamseeachanta/workspace-hub

Independently verify subagent-claimed file writes with filesystem and git checks before treating the artifact as real, before committing it, and before referencing the path in downstream prompts.

teams-meeting-pipeline

from vamseeachanta/workspace-hub

Operate the Teams meeting summary pipeline via Hermes CLI — summarize meetings, inspect pipeline status, replay jobs, manage Microsoft Graph subscriptions.

subagent-driven-development

from vamseeachanta/workspace-hub

Use when executing implementation plans with independent tasks. Dispatches fresh delegate_task per task with two-stage review (spec compliance then code quality).