Skip to content

FrancyJGLisboa/agent-skills-platform / references/examples/stock-analyzer

stock-analyzer

Unverified

Provides comprehensive technical analysis for stocks and ETFs using RSI, MACD, Bollinger Bands, and other indicators. Activates when user requests stock analysis, technical indicators, trading signals, or market data for specific ticker symbols.

evals/stock-analyzer.eval.md

View on GitHub
# Eval Spec: stock-analyzer

This skill ships its own loss function: a set of binary checks plus golden cases
that define what "the analysis worked" means. Run it anytime with
`python3 scripts/run_evals.py` (scores the golden baseline) or
`python3 scripts/run_evals.py --rollout` (runs the skill end-to-end on each golden
input and scores the real output).

## Criteria

1. **valid-json** (command) — the produced output parses as JSON.
2. **has-ticker** (command) — the output carries the requested `ticker`.
3. **has-signal-action** (command) — a trading `signal.action` is present
   (BUY / SELL / HOLD).
4. **signal-is-justified** (llm-judge) — the `signal.reasoning` actually supports
   the chosen action (not graded by the runner; checklist only).

## Golden cases

Three representative analysis requests. The skill embeds a timestamp in its output
(`datetime.now()`), so outputs are not byte-stable — the cases are seeded as
`pending-first-green`: the structural command checks above pass on any run, and you
can capture a concrete baseline with
`python3 scripts/run_evals.py --rollout --promote`.

- **case-1** — AAPL with RSI + MACD.
- **case-2** — MSFT with RSI + MACD + Bollinger.
- **case-3** — GOOGL with RSI only.

## Spec

```json
{
  "skill": "stock-analyzer",
  "run": "python3 scripts/run_pipeline.py --input {input} --output {output}",
  "criteria": [
    {"id": "valid-json", "text": "Output parses as JSON", "type": "command", "cmd": "python3 -c \"import json,sys; json.load(open(sys.argv[1]))\" {output}"},
    {"id": "has-ticker", "text": "Output carries a ticker", "type": "command", "cmd": "python3 -c \"import json,sys; d=json.load(open(sys.argv[1])); assert d.get('ticker')\" {output}"},
    {"id": "has-signal-action", "text": "Output has a signal.action", "type": "command", "cmd": "python3 -c \"import json,sys; d=json.load(open(sys.argv[1])); assert d['signal']['action'] in ('BUY','SELL','HOLD')\" {output}"},
    {"id": "signal-is-justified", "text": "signal.reasoning supports the chosen action", "type": "llm-judge"}
  ],
  "golden": [
    {"id": "case-1", "input": "golden/case-1/input.json", "expected": null, "split": "val", "expected_status": "pending-first-green", "compare_ignore": ["timestamp"]},
    {"id": "case-2", "input": "golden/case-2/input.json", "expected": null, "split": "test", "expected_status": "pending-first-green", "compare_ignore": ["timestamp"]},
    {"id": "case-3", "input": "golden/case-3/input.json", "expected": null, "split": "val", "expected_status": "pending-first-green", "compare_ignore": ["timestamp"]}
  ]
}
```
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall