feat(cardex): retune value model for spread + add card image renderer
Retune cardex against the live 24-VM library: the old rubric tied 11 VMs at 615 Rare (completeness/reusability maxed for any done+shareable VM, only heft varied). Rebalance the five sub-score caps (richness 20→30 becomes the discriminator via per-tag CAPABILITY_WEIGHT; heft log2-scaled) and recalibrate the rarity thresholds. Library now spreads 2 Legendary / 3 Epic / 16 Rare / 2 Uncommon / 2 Common instead of a single fat Rare band. Add cmd_chat/cardimg.py: render a Card as a collectible trading-card image. Free + local-first — a hand-built SVG frame (rarity holo, type badges, 6 stat bars, dex#, flavor) plus a deterministic procedural creature sigil that needs no network, rasterized to PNG via local cairosvg/ImageMagick. Pluggable art backends are opt-in: free hosted Pollinations flux, and key-gated paid Stability / OpenAI / RunwayML. Every backend only fills the portrait window, so a card always renders fully offline; failures fall back to the sigil. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
leetcrypt
+61
-22
@@ -32,12 +32,41 @@ BASE_MB = 128.0 # ~size of the empty hh-dev base image; payload = size
|
||||
SUBSCORE_MAX = 100 # completeness+reusability+richness+pedigree+heft caps sum
|
||||
POWER_MAX = 1000 # PowerScore = subscore_sum * 10
|
||||
|
||||
# Sub-score caps (sum == SUBSCORE_MAX). completeness/reusability are near-binary
|
||||
# "is it a real, shipped VM?" gates, so they're kept small; the discriminating
|
||||
# signal lives in richness (how self-describing + how rare its capabilities),
|
||||
# heft (installed payload), and pedigree — those carry most of the spread.
|
||||
COMP_MAX = 20.0
|
||||
REUSE_MAX = 15.0
|
||||
RICH_MAX = 30.0
|
||||
PED_MAX = 10.0
|
||||
HEFT_MAX = 25.0
|
||||
|
||||
# Per-tag capability weight — rare/high-skill security capabilities are worth
|
||||
# more than generic ones, so a VM's "type rarity" pushes its richness (and thus
|
||||
# its rarity tier). Unlisted tags default to 1.5.
|
||||
CAPABILITY_WEIGHT = {
|
||||
# marquee / rare offensive + IR capabilities
|
||||
"exploit": 3.0, "pwn": 3.0, "c2": 3.0, "payload": 3.0, "malware": 3.0,
|
||||
"crack": 3.0, "redteam": 3.0, "dfir": 3.0, "forensics": 3.0, "reverse": 3.0,
|
||||
"ransomware": 3.0,
|
||||
# solid mid-tier capabilities
|
||||
"recon": 2.0, "osint": 2.0, "fuzz": 2.0, "brute": 2.0, "crypto": 2.0,
|
||||
"tls": 2.0, "dns": 2.0, "yara": 2.0, "pcap": 2.0, "vuln": 2.0, "cve": 2.0,
|
||||
"ids": 2.0, "threat": 2.0, "nmap": 2.0,
|
||||
# generic / scaffolding tags
|
||||
"dev": 1.0, "build": 1.0, "lib": 1.0, "http": 1.0, "web": 1.0, "log": 1.0,
|
||||
"network": 1.0, "audit": 1.0, "blue": 1.0, "training": 1.0, "lab": 1.0,
|
||||
"security": 1.0, "smoke": 0.0, "slice": 0.0,
|
||||
}
|
||||
|
||||
# Rarity by absolute PowerScore cutoff (user choice: stable & reproducible).
|
||||
# Calibrated against the live library distribution so tiers actually spread.
|
||||
RARITY_TIERS = [
|
||||
("Legendary", 825),
|
||||
("Epic", 650),
|
||||
("Rare", 450),
|
||||
("Uncommon", 250),
|
||||
("Legendary", 720),
|
||||
("Epic", 660),
|
||||
("Rare", 540),
|
||||
("Uncommon", 380),
|
||||
("Common", 0),
|
||||
]
|
||||
|
||||
@@ -82,37 +111,47 @@ def _clamp(x: float, lo: float = 0.0, hi: float = 1.0) -> float:
|
||||
|
||||
# ── the five value sub-scores ────────────────────────────────────────────────
|
||||
def _completeness(status: str, todo: list, blockers: list) -> float:
|
||||
"""Is it a real, finished VM? Near-binary gate; small cap by design."""
|
||||
s = (status or "").strip().lower()
|
||||
base = 25.0 if s in ("done", "complete", "completed") else (12.0 if s == "in_progress" else 0.0)
|
||||
base = COMP_MAX if s in ("done", "complete", "completed") else (10.0 if s == "in_progress" else 0.0)
|
||||
base -= min(len(todo) * 3, 12)
|
||||
base -= len(blockers) * 5
|
||||
return _clamp(base, 0, 25)
|
||||
return _clamp(base, 0, COMP_MAX)
|
||||
|
||||
|
||||
def _reusability(shareable: bool, share_path: str, artifact_kind: str) -> float:
|
||||
v = 14.0 if (shareable and share_path) else 0.0
|
||||
v += {"file": 6.0, "image": 2.0, "snapshot": 2.0}.get(artifact_kind, 0.0)
|
||||
return _clamp(v, 0, 20)
|
||||
"""Can another agent actually pull and reuse it?"""
|
||||
v = 10.0 if (shareable and share_path) else 0.0
|
||||
v += {"file": 5.0, "image": 3.0, "snapshot": 2.0}.get(artifact_kind, 0.0)
|
||||
return _clamp(v, 0, REUSE_MAX)
|
||||
|
||||
|
||||
def _richness(purpose: str, objective: str, tags: list, has_usage: bool) -> float:
|
||||
"""The discriminating axis: how self-describing + how rare its capabilities.
|
||||
|
||||
Rewards prose depth (purpose/objective length), tag breadth, *and* the
|
||||
capability-rarity of those tags (a c2/exploit VM out-scores a dev-lib one)."""
|
||||
v = 0.0
|
||||
v += 5 if purpose.strip() else 0
|
||||
v += 5 if objective.strip() else 0
|
||||
v += min(len(tags) * 2, 6)
|
||||
v += 4 if has_usage else 0
|
||||
return _clamp(v, 0, 20)
|
||||
v += min(len(purpose.split()) * 0.8, 8) # purpose prose depth
|
||||
v += 5 if objective.strip() else 0 # has a stated objective
|
||||
v += min(len(tags) * 1.5, 9) # tag breadth
|
||||
cap = sum(CAPABILITY_WEIGHT.get(t.strip().lower(), 1.5) for t in tags)
|
||||
v += min(cap, 8) # capability rarity
|
||||
v += 4 if has_usage else 0 # documents how to run
|
||||
return _clamp(v, 0, RICH_MAX)
|
||||
|
||||
|
||||
def _pedigree(provenance: list, created_by: str) -> float:
|
||||
v = min(len(provenance) * 3, 12)
|
||||
v += 3 if created_by.strip().lower() not in ("", "smoke", "pulled") else 0
|
||||
return _clamp(v, 0, 15)
|
||||
"""How battle-tested / hand-built its lineage is."""
|
||||
v = min(len(provenance) * 3, 8)
|
||||
v += 2 if created_by.strip().lower() not in ("", "smoke", "pulled") else 0
|
||||
return _clamp(v, 0, PED_MAX)
|
||||
|
||||
|
||||
def _heft(size_bytes: int) -> float:
|
||||
"""Installed payload over the empty base — log-scaled so giants don't run away."""
|
||||
payload_mb = max(0.0, (size_bytes or 0) / 1e6 - BASE_MB)
|
||||
return _clamp(4.0 * math.log2(1 + payload_mb / 8.0), 0, 20)
|
||||
return _clamp(5.0 * math.log2(1 + payload_mb / 6.0), 0, HEFT_MAX)
|
||||
|
||||
|
||||
def _rarity_of(power: int) -> str:
|
||||
@@ -226,12 +265,12 @@ def card_from_entry(entry: dict, manifest: dict | None = None) -> Card:
|
||||
payload_mb = max(0.0, (entry.get("size_bytes") or 0) / 1e6 - BASE_MB)
|
||||
speed_raw = _clamp(1 - payload_mb / 256)
|
||||
stats = {
|
||||
"HP": _stat(floor + 0.70 * (0.6 * heft / 20 + 0.4 * comp / 25)),
|
||||
"HP": _stat(floor + 0.70 * (0.6 * heft / HEFT_MAX + 0.4 * comp / COMP_MAX)),
|
||||
"Attack": _stat(floor + 0.70 * off),
|
||||
"Defense": _stat(floor + 0.70 * (0.5 * dfn + 0.5 * comp / 25)),
|
||||
"SpAtk": _stat(floor + 0.70 * (rich / 20)),
|
||||
"Defense": _stat(floor + 0.70 * (0.5 * dfn + 0.5 * comp / COMP_MAX)),
|
||||
"SpAtk": _stat(floor + 0.70 * (rich / RICH_MAX)),
|
||||
"Speed": _stat(floor + 0.70 * speed_raw),
|
||||
"SpDef": _stat(floor + 0.70 * (0.6 * reuse / 20 + 0.4 * ped / 15)),
|
||||
"SpDef": _stat(floor + 0.70 * (0.6 * reuse / REUSE_MAX + 0.4 * ped / PED_MAX)),
|
||||
}
|
||||
|
||||
shiny = (_seed(label) >> 16) % 16 == 0 # ~1/16, the classic shiny odds
|
||||
|
||||
Reference in New Issue
Block a user