.claude/skills/analyze-context/SKILL.md
# Analyze Context Analyzes a gameplay session context dump from Gradient Bang. Summarizes agent activity and, if the user reports a problem, identifies where it occurred and scans source code for the root cause. ## Parameters - **context** (required): The gameplay session context dump (paste inline or provide a file path). - **problem** (optional): A description of what went wrong during the session. ## Steps ### 1. Receive input The user provides a context dump, either inline in the messa
npx skillsauth add pipecat-ai/gradient-bang .claude/skills/analyze-contextInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Analyzes a gameplay session context dump from Gradient Bang. Summarizes agent activity and, if the user reports a problem, identifies where it occurred and scans source code for the root cause.
The user provides a context dump, either inline in the message or as a file path. If they provide a file path, read the file. If neither is provided, check whether there is a currently open file in the IDE (it will appear in the conversation context tagged as ide_opened_file). If there is, read that file and use it as the context dump. If there is no open file either, ask:
"Please paste your context dump or provide a file path."
"Did anything go wrong during this session, or do you just want a summary?"
Scan the full context dump and produce a terse, factual summary:
Keep this section short — bullet points preferred, no padding.
Scan the context for evidence of the problem:
Using the issue location and agent identified in Step 4, read the relevant source files:
src/gradientbang/pipecat_server/bot.py — pipeline wiring, MainAgentsrc/gradientbang/pipecat_server/subagents/voice_agent.pysrc/gradientbang/pipecat_server/subagents/task_agent.pysrc/gradientbang/pipecat_server/subagents/event_relay.pysrc/gradientbang/pipecat_server/subagents/ui_agent.pysrc/gradientbang/tools/ — tool schemasFocus on:
Cite file paths and line numbers for every suspect location.
Produce the final report in exactly this format:
## Summary of session
<terse bullet-point summary of VoiceAgent and TaskAgent activity>
## Cited issue(s)
<quoted excerpt(s) from the context showing where the problem occurred, with agent attribution and context position>
<file:line references from source code identified as the likely cause>
## Potential fixes
<concrete, specific suggested code changes with file:line references>
If no problem was reported and the user confirmed they just want a summary, omit Cited issue(s) and Potential fixes entirely.
src/gradientbang/tools/ and are shared across VoiceAgent, TaskAgent, and UIAgent — check there when a tool call looks malformed.testing
# Restore Ship Restores a ship to a healthy state for testing. Clears the `destroyed_at` flag, restocks fighters/shields/warp power to max values from the ship definition, re-adds corporation ships to `corporation_ships` if missing, and recreates the pseudo-character record for corp ships so tasks can be issued. No events are emitted — this is a database-only operation. ## Parameters Ask the user for: - **ship_id**: UUID of the ship to restore (required) ## Steps ### 1. Source environment
development
# Reset World Resets the game database, generates a fresh universe, loads quest definitions, and seeds combat cron config. ## Parameters The user specifies the environment as an argument: `/reset-world local`, `/reset-world dev`, or `/reset-world prod`. If not provided, ask which environment. - `local` → env file: `.env.supabase` - `dev` → env file: `.env.cloud.dev` - `prod` → env file: `.env.cloud` Additional optional parameters (ask if not provided, or use defaults): - **Sector count**: n
devops
# NPC Runs an autonomous AI task agent as a game character. Resolves a character name to its UUID, then launches the `npc-run` script which connects to the game server and executes the given task using a Pipecat + Gemini LLM pipeline. ## Parameters - **character_name** (required): The character's display name (e.g. `JOETRADER`). Passed as the argument to `/npc`. - **task** (required): A natural language description of what the character should do (e.g. "Explore and find 5 new sectors"). If `
data-ai
# Run Database Migration Applies pending Supabase migrations to the local or production database. This skill is strictly additive — it only applies new migrations and NEVER resets, truncates, or drops existing data. ## Safety rules — READ BEFORE PROCEEDING **You MUST follow every rule below. No exceptions.** 1. **NEVER** run `supabase db reset`, `supabase db reset --linked`, or any command that drops/recreates the database. 2. **NEVER** run `DROP TABLE`, `DROP SCHEMA`, `TRUNCATE`, or `DELETE