Agent Claude Code
Use when reverse-engineering or implementing anything in customcomponents/haggle/agl/ — the Auth0 token lifecycle (AglAuth), the REST API calls (AglClient), or the JSON response parsers. Also the go-to agent when debugging AGL HTTP behaviour, 401s, rate limits, or response shape questions. Owns all files under…
Agent Claude Code
Use proactively when editing coordinator.py, agl/client.py, or any async function in customcomponents/haggle/. Reviews for blocking I/O in the HA event loop (the integration's.
Agent Claude Code
Use proactively after non-trivial edits in customcomponents/haggle/ (excluding pure auth/perf-critical files owned by other agents). Reviews for project-specific quality rules and verifies tooling is clean. Defers to /simplify skill for general simplification.
Agent Claude Code
Use when defining or reviewing sensor stateclass, deviceclass, unitofmeasurement, or anything that feeds the HA Energy dashboard. Also owns importstatistics() usage, NEM12 semantics, and tariff/cost calculation logic. Call this agent before shipping any energy or monetary sensor.
Agent Claude Code
Use proactively when designing or modifying anything in customcomponents/haggle/ that touches HA core APIs: init.py, configflow.py, coordinator.py, sensor.py, entity base classes, runtimedata, or platform setup. Confirms HA 2025.x+ patterns and catches regressions against HA quality scale requirements.
Agent Claude Code
Use proactively after any code change in customcomponents/haggle/ to write or update tests. Owns all files under tests/. Uses pytest-homeassistant-custom-component patterns and the anonymised AGL API fixtures in tests/fixtures/.
Agent Claude Code
Use only when cutting a release via the /release command. Bumps manifest.json version, updates CHANGELOG.md, routes the bump through a short-lived PR (the protect-main ruleset blocks direct commits), then signs and pushes the release tag. Refuses to run if the working tree is dirty or CI is not green.
Agent Claude Code
Use proactively when editing configflow.py, agl/, init.py, or any code touching tokens/auth/HTTP. Reviews for credential leakage, OAuth/PKCE correctness, token rotation persistence, and dependency vulnerabilities. Project-scoped sibling of the built-in /security-review skill.
Command Claude Code
Validate the integration against Home Assistant's hassfest rules.
Command Claude Code
Scaffold a new sensor entity with tests.
Command Claude Code
Push the current feature branch and open a GitHub pull request, after running a full documentation audit.
Command Claude Code
Command "release" from NaanyaBiz/haggle, covering /release, usage, pre-conditions (checked before proceeding), delegates to and after completion.
Command Claude Code
Manage sibling git worktrees for parallel feature development.
Hook Claude Code
Runs after a tool call finishes for Edit, Write, MultiEdit, Edit and Write tool calls, executing format-python.sh and validate-manifest.sh (2 commands). From NaanyaBiz/haggle.
Hook Claude Code
Runs before the agent uses a tool for Bash tool calls, executing guard-main-branch.sh. From NaanyaBiz/haggle.
Hook Claude Code
Runs when you submit a prompt, before the agent sees it, executing inject-branch-context.sh. From NaanyaBiz/haggle.
Settings file Claude Code
Agent settings declaring 3 hook events (PostToolUse, PreToolUse, UserPromptSubmit) and 26 allowed tools.
Instructions file CodexOpenCode
AGENTS.md instructions for NaanyaBiz/haggle, covering agents.md — haggle integration guide, dev loop, install deps (once, or after pyproject.toml changes), run tests and lint + format.
Instructions file
Claude Code instructions for NaanyaBiz/haggle, a project described as: AGL Australia smart meter -> Home Assistant Energy dashboard (Claude-generated custom integration).
Agent
A fixed set of hostile issue/PR/attachment payloads used to sanity-check the haggle-triage routine's UNTRUSTED-CONTENT ARMOUR (triage-routine.md) before any change to its prompt.
Agent
Agent "triage-routine" from NaanyaBiz/haggle, covering agent: haggle-triage, change control (repo-first), configuration, what it does and what it never does.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: