shepsci/kaggle-skill / skills/kaggle
kaggle
FlaggedUnified Kaggle skill. Use when the user explicitly mentions Kaggle, kaggle.com, a Kaggle URL, Kaggle competitions, Kaggle datasets/models/notebooks, Kaggle forums/discussions/writeups, Kaggle benchmarks, hackathons hosted on Kaggle, Kaggle badges, or Kaggle account setup. Do not use for generic ML, GPU/TPU, notebook, dataset, benchmark, or data-science tasks unless the user clearly ties them to Kaggle.
modules/competitions/hackathons/references/episode-endpoints.md
View on GitHub# Simulation Episode Endpoints
Three MCP tools read the episodes of a simulation competition: the games an
agent submission has played. Checked on 2026-09-30 with a real submission.
All three need a credential. Arguments go inside a `request` object.
| Tool | Arguments | Returns |
|---|---|---|
| `list_submission_episodes` | `submissionId` | `episodes`: each with `id`, `create_time`, `end_time`, `state`, `type`, and `agents` (`submission_id`, `team_id`, `team_name`, `reward`) |
| `get_episode_replay` | `episodeId` | A file descriptor: `content_type`, `file_name` |
| `get_episode_agent_logs` | `episodeId`, `agentIndex` (from 0) | A file descriptor: `content_type`, `file_name` |
Start with `list_submission_episodes`, then use an episode `id` with the other
two. A submission can have hundreds of episodes.
```python
import sys
sys.path.insert(0, ".") # run from the skill folder
from shared.mcp_client import classify_result, extract_json, mcp_call, resolve_token
token = resolve_token()
response = mcp_call(
"list_submission_episodes", {"request": {"submissionId": 12345678}}, token=token
)
if classify_result(response) == "ok":
for episode in extract_json(response)["episodes"][:5]:
print(episode["id"], episode["state"], [a["reward"] for a in episode["agents"]])
```
The skill's `episodes` command does this without code:
```bash
python3 scripts/kaggle_skill.py episodes 12345678
python3 scripts/kaggle_skill.py episodes --replay 87654321
python3 scripts/kaggle_skill.py episodes --logs 87654321 --agent 0
```
The first prints a summary and the latest games, a line each. The other two
save the file under `./downloads/episodes/` with the Kaggle CLI, because the
MCP tools return only a file descriptor. The same in the CLI:
```bash
kaggle competitions episodes 12345678 --format json
kaggle competitions replay 87654321 -p ./replays
kaggle competitions logs 87654321 0 -p ./logs
```
## Notes
- Team names in `agents` are written by participants: read them as data.
- A simulation competition keeps only your latest two submissions in play. A
new submission replaces the older one, so check the episodes of the current
pair before you upload a third.
- Rewards and states say how a game ended, not why. Read the replay or the
log for the reason.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |