docs(polymarket-browse): remove Changelog and Credits from SKILL.md

Changelog tracked via git tags instead.

skills/polymarket-browse/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: polymarket-browse
+version: 0.0.3
 category: research
 description: Browse tradeable Polymarket events by game category. Shows active matches with ML odds (cents format), volume, tournament, and market URLs. Supports Counter Strike, League of Legends, Dota 2, Valorant, NBA, NFL, UFC, Tennis.
 ---
@@ -34,7 +35,7 @@ hermes mcp add polymarket https://docs.polymarket.com/mcp
 ## Usage
 
 ```
-polymarket-browse [--category "Counter Strike"] [--limit 5] [--matches N] [--non-matches N] [--search "TeamName"] [--matches-only] [--non-matches-only] [--detail N] [--raw] [--telegram] [--no-cache] [--max-total N]
+polymarket-browse [--category "Counter Strike"] [--limit 5] [--matches N] [--non-matches N] [--search "TeamName"] [--matches-only] [--non-matches-only] [--detail N] [--raw] [--telegram] [--no-cache] [--max-total N] [--starts-before TIMESTAMP] [--timezone UTC+X]
 ```
 
 ## Arguments
@@ -46,11 +47,17 @@ polymarket-browse [--category "Counter Strike"] [--limit 5] [--matches N] [--non
 - `--search` : Free-text team/term search within the selected category. Appends to the category query. Example: `--category "Counter Strike" --search "FlyQuest"`
 - `--matches-only` : Show only match markets (suppress non-match section).
 - `--non-matches-only` : Show only non-match markets (suppress match section).
-- `--detail` : Index of match event (1-indexed) to show detailed markets. Default: 1. Set to 0 to disable.
+- `--detail N` : Show detailed markets for match event N (1-indexed).
+  - Only applies to MATCH markets (not non-match/tournament markets)
+  - Default: 1 (auto-shows details for first match)
+  - Set to 0 to disable detail view
+  - If N exceeds available matches, shows error with available range
 - `--list-categories` : List available game categories and exit
 - `--raw` : Show all events without tradeable filter (for debugging). Includes fetch stats.
 - `--no-cache` : Disable caching and fetch fresh data from the API.
 - `--max-total` : Maximum total events to fetch before early exit. Default: no limit. Useful for quick snapshots.
+- `--starts-before` : Unix timestamp filter. Only show match events starting before this time (LIVE events always shown regardless of timestamp).
+- `--timezone` : Timezone for displaying times. Format: `UTC+X` or `UTC-X` (e.g., `UTC+7`, `UTC-5`). Default: UTC+7 (WIB).
 - `--telegram` : Send results to Telegram. Requires `BOT_TOKEN` and `CHAT_ID` in environment variables.
 
 ## Output Format
@@ -135,9 +142,12 @@ The script first fetches page 1 to determine total pages, then fetches remaining
 
 ## Rate Limiting
 
-- Exponential backoff: 2s → 4s → 8s → 16s → 32s
+- TokenBucket rate limiter: 10 API calls per second
+- Exponential backoff on retries: 2s → 4s → 8s → 16s → 32s
 - Max 5 retries before aborting
 
+**URL Encoding**: Special characters in `--search` (e.g., `&`, `=`, `%`, `+`, `#`) are properly encoded to prevent URL injection.
+
 ## Caching
 
 Results are cached in `~/.cache/polymarket-browse/` with a **5-minute TTL** to reduce redundant API calls.
@@ -152,3 +162,91 @@ All odds are shown in **cents** format:
 - `30c` = 0.30 probability
 - `95c` = 0.95 probability
 - `GamerLegion 28c | 72c Team Yandex` = GamerLegion at 28c, Team Yandex at 72c
+
+## Troubleshooting
+
+### "WARNING: Partial fetch" appears
+The API returned incomplete data due to an error/timeout. Results shown may be incomplete. Try again with `--no-cache` to force a fresh fetch.
+
+### No markets appear
+- Verify your category is correct: `--list-categories`
+- Try with `--raw` to see all events (not just tradeable ones)
+- Some categories may have no active match markets at certain times
+
+### Why did my match disappear?
+Matches are filtered out when:
+- They have ended (startTime > 4 hours ago)
+- BO2 matches ended in a tie (1-1)
+- The market has converged (bestBid >= 0.99 or bestAsk <= 0.01)
+- The event has ended (endDate passed)
+
+### Telegram not working
+- Verify `BOT_TOKEN` and `CHAT_ID` environment variables are set
+- Ensure bot is started and chat ID is correct
+- Check Telegram has not blocked the bot
+
+## Examples
+
+### Basic usage
+```bash
+# Browse Counter Strike matches (default)
+polymarket-browse
+
+# Browse NBA matches
+polymarket-browse --category NBA
+
+# Show more results
+polymarket-browse --limit 10
+```
+
+### Searching for teams
+```bash
+# Find FlyQuest Counter Strike matches
+polymarket-browse --category "Counter Strike" --search "FlyQuest"
+
+# Find any team/event across category
+polymarket-browse --category "Counter Strike" --search "Spirit"
+```
+
+### Filtering results
+```bash
+# Show only match markets (no tournament futures)
+polymarket-browse --matches-only
+
+# Show only non-match markets (tournaments, props)
+polymarket-browse --non-matches-only
+
+# Different limits for each section
+polymarket-browse --matches 10 --non-matches 5
+```
+
+### Using --detail
+```bash
+# Show details for 1st match (default behavior, auto-enabled)
+polymarket-browse --detail 1
+
+# Show details for 3rd match
+polymarket-browse --detail 3
+
+# Disable detail view
+polymarket-browse --detail 0
+```
+
+### Debugging
+```bash
+# Show all events without tradeable filter
+polymarket-browse --raw
+
+# Force fresh data (bypass cache)
+polymarket-browse --no-cache
+
+# Limit total events for quick snapshot
+polymarket-browse --max-total 20
+```
+
+### Timezone
+```bash
+# Display times in different timezone (default: UTC+7/WIB)
+polymarket-browse --timezone UTC+8
+polymarket-browse --timezone UTC-5
+```

skills/polymarket-browse/scripts/browse.py

@@ -6,14 +6,16 @@ Browse tradeable Polymarket events by game category.
 
 import html
 import json
+import sys
 import time
 import argparse
 import hashlib
 import os
+import threading
 from concurrent.futures import ThreadPoolExecutor, as_completed
 from datetime import datetime, timezone, timedelta
 from typing import Any, Callable, TypedDict
-from urllib.parse import urlencode
+from urllib.parse import urlencode, quote
 from urllib.request import urlopen, Request
 
 
@@ -97,7 +99,93 @@ class FetchResult(TypedDict):
 PAGE_SIZE = 50
 MAX_RETRIES = 5
 INITIAL_RETRY_DELAY = 2  # exponential backoff starts at 2s
+MAX_RESPONSE_SIZE_MULTIPLIER = 10  # Response size limit = PAGE_SIZE * multiplier
+MAX_RESPONSE_SIZE_MIN = 10 * 1024 * 1024  # 10MB minimum
+MAX_RESPONSE_SIZE_MAX = 100 * 1024 * 1024  # 100MB maximum for safety
+RATE_LIMIT_CALLS = 10  # max API calls
+RATE_LIMIT_WINDOW = 1.0  # per second
 WIB = timezone(timedelta(hours=7))  # UTC+7 for Indonesian users
+_DISPLAY_TZ = WIB  # Module-level timezone for display (configurable via --timezone)
+
+
+def parse_timezone(tz_str: str) -> timezone:
+    """
+    Parse timezone string to datetime.timezone.
+    Supports: UTC offset format (UTC+7, UTC-5).
+    Falls back to WIB (UTC+7) on parse failure.
+    """
+    tz_str = tz_str.strip()
+    if tz_str.startswith("UTC"):
+        offset_str = tz_str[3:].strip()
+        if not offset_str:
+            return timezone.utc
+        sign = -1 if offset_str[0] == "-" else 1
+        if offset_str[0] in "+-":
+            offset_str = offset_str[1:]
+        try:
+            if ":" in offset_str:
+                hours, minutes = offset_str.split(":")
+                hours = int(hours)
+                minutes = int(minutes)
+            else:
+                hours = int(offset_str)
+                minutes = 0
+            total_minutes = hours * 60 + minutes
+            if sign == -1:
+                total_minutes = -total_minutes
+            return timezone(timedelta(minutes=total_minutes))
+        except ValueError:
+            return WIB
+    return WIB
+    try:
+        from datetime import ZoneInfo
+
+        return ZoneInfo(tz_str).utcoffset(None)
+    except Exception:
+        return WIB
+
+
+def get_max_response_size(page_size: int = PAGE_SIZE) -> int:
+    """
+    Calculate max response size based on expected payload.
+    Uses 10x multiplier: if PAGE_SIZE=50 events, expected ~500KB-5MB,
+    so 10x gives 5MB-50MB. Clamped between 10MB and 100MB.
+    """
+    multiplier = MAX_RESPONSE_SIZE_MULTIPLIER * page_size * 1024  # rough estimate
+    size = max(multiplier, MAX_RESPONSE_SIZE_MIN)
+    return min(size, MAX_RESPONSE_SIZE_MAX)
+
+
+class RateLimiter:
+    """Token bucket rate limiter for API calls. Thread-safe for use with ThreadPoolExecutor."""
+
+    def __init__(
+        self, calls: int = RATE_LIMIT_CALLS, window: float = RATE_LIMIT_WINDOW
+    ):
+        self.calls = calls
+        self.window = window
+        self.tokens = float(calls)
+        self.last_update = time.monotonic()
+        self._lock = threading.Lock()
+
+    def acquire(self) -> None:
+        """Block until a token is available."""
+        with self._lock:
+            now = time.monotonic()
+            elapsed = now - self.last_update
+            self.tokens = min(
+                self.calls, self.tokens + elapsed * (self.calls / self.window)
+            )
+            if self.tokens < 1:
+                wait_time = (1 - self.tokens) * (self.window / self.calls)
+                time.sleep(wait_time)
+                self.tokens = 0
+            else:
+                self.tokens -= 1
+            self.last_update = time.monotonic()
+
+
+_rate_limiter = RateLimiter()
 
 GAME_CATEGORIES = {
     "All Esports": "Esports",
@@ -166,7 +254,7 @@ def fetch_page(
 ) -> dict[str, Any] | None:
     base = "https://gamma-api.polymarket.com/public-search"
     url = (
-        f"{base}?q={q.replace(' ', '%20')}&limit={PAGE_SIZE}&page={page}"
+        f"{base}?q={quote(q, safe='')}&limit={PAGE_SIZE}&page={page}"
         f"&search_profiles=false&search_tags=false"
         f"&keep_closed_markets=0&events_status=active&cache=false"
     )
@@ -176,9 +264,16 @@ def fetch_page(
         if attempt > 0:
             time.sleep(delay)
         try:
+            _rate_limiter.acquire()
             req = Request(url, headers={"User-Agent": "Mozilla/5.0"})
             with urlopen(req, timeout=10) as r:
-                return json.loads(r.read())
+                data = r.read()
+                max_size = get_max_response_size(PAGE_SIZE)
+                if len(data) > max_size:
+                    raise ValueError(
+                        f"API response too large: {len(data)} bytes (max {max_size})"
+                    )
+                return json.loads(data)
         except Exception:
             if attempt < max_retries - 1:
                 delay *= 2
@@ -383,7 +478,7 @@ def is_tradeable_event(e: dict[str, Any]) -> bool:
             now = datetime.now(timezone.utc)
             if end_dt < now:
                 return False
-        except:
+        except (ValueError, TypeError):
             pass
 
     # Filter: match has already started (startTime is in the past)
@@ -397,7 +492,7 @@ def is_tradeable_event(e: dict[str, Any]) -> bool:
                 hours_ago = (now - start_dt).total_seconds() / 3600
                 if hours_ago > 4:
                     return False
-        except:
+        except (ValueError, TypeError):
             pass
 
     return True
@@ -453,12 +548,12 @@ def _get_time_data(e: dict[str, Any], tz: timezone | None = None) -> TimeData:
     Args:
         e: Event dict with 'startTime' or 'startDate' key.
         tz: datetime.timezone for abs_time formatting.
-            Defaults to WIB (UTC+7).
+            Defaults to _DISPLAY_TZ (set via --timezone, or WIB).
 
     Returns:
         TimeData with time_status, time_urgency, and abs_time
     """
-    tz = tz or WIB
+    tz = tz or _DISPLAY_TZ
     start_str = e.get("startTime") or e.get("startDate", "")
 
     if not start_str:
@@ -544,6 +639,47 @@ def sort_events(events: list[dict[str, Any]]) -> list[dict[str, Any]]:
 # ============================================================
 
 
+def _is_live_event(e: dict[str, Any]) -> bool:
+    """Check if event is LIVE (started within last 4 hours)."""
+    start_str = e.get("startTime") or e.get("startDate", "")
+    if not start_str:
+        return False
+    try:
+        start_dt = datetime.fromisoformat(start_str.replace("Z", "+00:00"))
+        now = datetime.now(timezone.utc)
+        delta = now - start_dt
+        if delta.total_seconds() < 0:
+            return False
+        hours_ago = delta.total_seconds() / 3600
+        return hours_ago < 4
+    except Exception:
+        return False
+
+
+def filter_by_starts_before(
+    events: list[dict[str, Any]], timestamp: int | None
+) -> list[dict[str, Any]]:
+    """Filter events to only include those starting before timestamp or LIVE events."""
+    if timestamp is None:
+        return events
+    filtered = []
+    for e in events:
+        start_str = e.get("startTime") or e.get("startDate", "")
+        if not start_str:
+            filtered.append(e)
+            continue
+        try:
+            start_dt = datetime.fromisoformat(start_str.replace("Z", "+00:00"))
+            start_ts = start_dt.timestamp()
+            if start_ts <= timestamp:
+                filtered.append(e)
+            elif _is_live_event(e):
+                filtered.append(e)
+        except Exception:
+            filtered.append(e)
+    return filtered
+
+
 def browse_events(
     q: str,
     matches_max: int = 10,
@@ -552,6 +688,7 @@ def browse_events(
     sort_by: str | None = None,
     max_total: int | None = None,
     use_cache: bool = True,
+    starts_before: int | None = None,
 ) -> BrowseResult:
     """
     Browse Polymarket events.
@@ -564,6 +701,7 @@ def browse_events(
         sort_by: None (fast, API order) or "volume" (full fetch, sort by volume desc)
         max_total: max total events to fetch before early exit (None = no limit)
         use_cache: whether to use cache (default True)
+        starts_before: unix timestamp filter for match events (None = no filter)
     """
     use_early_exit = sort_by is None
     fetch_matches_max = matches_max if use_early_exit else None
@@ -579,7 +717,8 @@ def browse_events(
     events = result["events"]
     match_events, non_match_events = filter_events(events, tradeable_only)
 
-    # Sort if requested; otherwise preserve API order
+    match_events = filter_by_starts_before(match_events, starts_before)
+
     if sort_by == "volume":
         match_events = sort_events(match_events)
         non_match_events = sort_events(non_match_events)
@@ -819,11 +958,10 @@ def format_detail_event(e: dict[str, Any]) -> DetailEvent:
 
 
 def get_header_date() -> str:
-    """Return current date string like 'Mar 25, 2026'"""
+    """Return current date string like 'Mar 25, 2026' in display timezone."""
     now_utc = datetime.now(timezone.utc)
-    utc7 = timezone(timedelta(hours=7))
-    now_utc7 = now_utc.astimezone(utc7)
-    return now_utc7.strftime("%b %d, %Y")
+    now_display = now_utc.astimezone(_DISPLAY_TZ)
+    return now_display.strftime("%b %d, %Y")
 
 
 def get_tournament(title: str) -> str:
@@ -1174,6 +1312,18 @@ def main() -> None:
         default=None,
         help="Max total events to fetch before early exit. Default: no limit.",
     )
+    parser.add_argument(
+        "--starts-before",
+        type=int,
+        default=None,
+        help="Unix timestamp filter. Only show match events starting before this time (LIVE events always shown).",
+    )
+    parser.add_argument(
+        "--timezone",
+        type=str,
+        default="UTC+7",
+        help="Timezone for displaying times (e.g., UTC+7, UTC-5). Default: UTC+7",
+    )
     parser.add_argument(
         "--telegram",
         action="store_true",
@@ -1193,6 +1343,9 @@ def main() -> None:
     matches_max = args.matches if args.matches is not None else args.limit
     non_matches_max = args.non_matches if args.non_matches is not None else args.limit
 
+    global _DISPLAY_TZ
+    _DISPLAY_TZ = parse_timezone(args.timezone)
+
     if args.search:
         print(f"\nFetching {args.category} events matching '{args.search}'...")
     else:
@@ -1205,6 +1358,7 @@ def main() -> None:
         tradeable_only=tradeable_only,
         max_total=args.max_total,
         use_cache=not args.no_cache,
+        starts_before=args.starts_before,
     )
 
     print_browse(
@@ -1224,10 +1378,15 @@ def main() -> None:
 
     # Print detail for selected event if any
     if result["match_events"] and args.detail > 0:
-        print("\n")
         idx = args.detail - 1
-        if idx < 0 or idx >= len(result["match_events"]):
-            idx = 0
+        num_events = len(result["match_events"])
+        if idx < 0 or idx >= num_events:
+            print(
+                f"Error: --detail {args.detail} is out of range (available: 1-{num_events}).",
+                file=sys.stderr,
+            )
+            sys.exit(1)
+        print("\n")
         detail_event = result["match_events"][idx]
         detail = format_detail_event(detail_event)
         print_detail(detail_event, detail)

skills/polymarket-browse/tests/test_browse.py

@@ -1839,5 +1839,155 @@ class TestBrowseEvents(unittest.TestCase):
         self.assertIn("partial", result)
 
 
+class TestStartsBeforeFilter(unittest.TestCase):
+    """Tests for --starts-before filter in browse_events()."""
+
+    def _make_event(self, event_id, start_time, volume="50000"):
+        """Helper to create a minimal match event with startTime and valid tradeable data."""
+        return {
+            "id": event_id,
+            "title": f"Match {event_id}",
+            "seriesSlug": "x",
+            "gameId": "1",
+            "startTime": start_time,
+            "markets": [
+                {
+                    "sportsMarketType": "moneyline",
+                    "volume": volume,
+                    "bestBid": "0.50",
+                    "bestAsk": "0.52",
+                    "acceptingOrders": True,
+                    "closed": False,
+                }
+            ],
+        }
+
+    @patch("browse.fetch_all_pages")
+    def test_starts_before_filters_future_events(self, mock_fetch):
+        """Events with startTime > timestamp should be filtered out."""
+        from browse import browse_events
+
+        mock_fetch.return_value = {
+            "events": [
+                self._make_event(
+                    "m1", "2026-03-27T14:00:00Z"
+                ),  # After cutoff (14:00 > 12:00)
+                self._make_event("m2", "2026-03-28T12:00:00Z"),  # After cutoff
+            ],
+            "total_raw": 2,
+            "partial": False,
+        }
+
+        # 2026-03-27T12:00:00Z = 1774612800
+        result = browse_events("test", starts_before=1774612800)
+
+        self.assertEqual(len(result["match_events"]), 0)
+
+    @patch("browse.fetch_all_pages")
+    def test_starts_before_includes_past_events(self, mock_fetch):
+        """Events with startTime <= timestamp should be included."""
+        from browse import browse_events
+
+        mock_fetch.return_value = {
+            "events": [
+                self._make_event(
+                    "m1", "2026-03-27T10:00:00Z"
+                ),  # Before cutoff (10:00 < 12:00)
+                self._make_event(
+                    "m2", "2026-03-27T11:00:00Z"
+                ),  # Before cutoff (11:00 < 12:00)
+            ],
+            "total_raw": 2,
+            "partial": False,
+        }
+
+        # 2026-03-27T12:00:00Z = 1774612800
+        result = browse_events("test", starts_before=1774612800)
+
+        self.assertEqual(len(result["match_events"]), 2)
+
+    @patch("browse.fetch_all_pages")
+    def test_starts_before_without_timestamp(self, mock_fetch):
+        """Without starts_before, all events should be returned."""
+        from browse import browse_events
+
+        mock_fetch.return_value = {
+            "events": [
+                self._make_event("m1", "2026-03-27T14:00:00Z"),
+                self._make_event("m2", "2026-03-28T12:00:00Z"),
+            ],
+            "total_raw": 2,
+            "partial": False,
+        }
+
+        result = browse_events("test")
+
+        # No filter, all events returned
+        self.assertEqual(len(result["match_events"]), 2)
+
+
+class TestTimezoneParsing(unittest.TestCase):
+    """Tests for parse_timezone() and timezone display."""
+
+    def test_parse_timezone_utc_plus7(self):
+        """UTC+7 should parse to WIB."""
+        from browse import parse_timezone
+        from datetime import timezone, timedelta
+
+        tz = parse_timezone("UTC+7")
+        self.assertEqual(tz, timezone(timedelta(hours=7)))
+
+    def test_parse_timezone_utc_minus5(self):
+        """UTC-5 should parse correctly."""
+        from browse import parse_timezone
+        from datetime import timezone, timedelta
+
+        tz = parse_timezone("UTC-5")
+        self.assertEqual(tz, timezone(timedelta(hours=-5)))
+
+    def test_parse_timezone_utc_no_offset(self):
+        """UTC should return timezone.utc."""
+        from browse import parse_timezone
+
+        tz = parse_timezone("UTC")
+        self.assertEqual(tz, timezone.utc)
+
+    def test_parse_timezone_with_minutes(self):
+        """UTC+5:30 should parse correctly."""
+        from browse import parse_timezone
+        from datetime import timezone, timedelta
+
+        tz = parse_timezone("UTC+5:30")
+        self.assertEqual(tz, timezone(timedelta(hours=5, minutes=30)))
+
+    def test_parse_timezone_invalid_falls_back_to_wib(self):
+        """Invalid timezone should fall back to WIB."""
+        from browse import parse_timezone
+        from datetime import timezone, timedelta
+
+        tz = parse_timezone("Invalid/Timezone")
+        self.assertEqual(tz, timezone(timedelta(hours=7)))
+
+
+class TestUrlEncoding(unittest.TestCase):
+    """Tests for proper URL encoding of search queries."""
+
+    def test_quote_encodes_special_chars(self):
+        """quote() should properly encode all special characters."""
+        from urllib.parse import quote
+
+        test_cases = [
+            ("Team A", "Team%20A"),
+            ("Team A & Team B", "Team%20A%20%26%20Team%20B"),
+            ("a=b", "a%3Db"),
+            ("100%", "100%25"),
+            ("C++", "C%2B%2B"),
+            ("Team (A)", "Team%20%28A%29"),
+            ("Team#1", "Team%231"),
+        ]
+        for input_str, expected in test_cases:
+            self.assertEqual(quote(input_str, safe=""), expected)
+
+
 if __name__ == "__main__":
     unittest.main()