feat(sn60): add Dexterity104-20260723-01

[Dexterity104]

Adds one miner bundle for the sn60__bitsec miner lane.

[carlos4s]

Thanks for the PR. I’m pausing this one before it enters a challenge.

Short version: the miner looks too benchmark-shaped. The bot overlords are suspicious, and this time they brought receipts.

What tripped the wire:

• agent.py: Screening review required: submission agent is highly similar to the current lane king implementation (similarity 0.94).

LLM sanity check:

• Reason: I do not see benchmark hardcoding or secret leakage in the submitted code, and it behaves like a generic repo auditor. However, the deterministic similarity score to the king agent is very high, so this should get a manual copycat review before acceptance.
• Verdict: suspicious. Confidence: 62% medium - plausible concern; manual review needed.
• Audit artifact saved for maintainer review.
• Proof: File header describes a generic multi-language whole-repo auditor, which is allowed and does not itself show benchmark replay. (agent.py:1)
• Proof: Uses validator-provided inference proxy and environment key x-inference-api-key, this is normal proxy usage and not an answer-key leak token. (agent.py:72)
• Proof: Large system prompt and checklist are generic audit instructions covering common bug classes across languages, not benchmark-specific findings or project IDs. (agent.py:96)

Why this matters: Kata agents need to analyze unseen projects. If the code recognizes benchmark fingerprints and spits out canned findings, that is replay, not mining.

How to fix it: remove project/finding-specific branches, static report banks, copied titles, and benchmark-only pattern matches. Keep reusable detectors and let the agent reason from the project source.

Maintainer commands: /kata approve accepts it, /kata review re-runs screening, and /kata close closes it.
You can also push an update and the checks will run again.