I build evidence gates for AI and security workflows.

The pattern is simple: take a claim, run it through a reproducible harness, and promote it only when the evidence holds. Proof before ship. If it cannot be rerun, reviewed, or signed, it is not done yet.

ANCHOR is the flagship because it has the clearest audience and the most finished artifact: a local-first proof gate for authorized smart-contract security research.

It turns raw tool signal into a structured case and keeps the full trail:

• seeded false claim -> rejected with REJECTED - INSUFFICIENT EVIDENCE
• FIXTURE CORPUS - REAL TOOL EXECUTION -> reproducible specimen run
• REPRODUCED_REAL -> a Foundry proof actually runs against the target
• NEGATIVE_CONTROL_PASSED -> the control stayed clean without pretending it was tested

ANCHOR is built to leave behind a reviewable artifact, not just a verdict:

• source
• evidence
• failed attempts
• successful reproduction
• remediation

That is the public story I want to be known for: signal to evidence, not signal to hype.

Visible repo:

• ANCHOR

These projects are all variations of the same core idea, applied in different places:

• AI-Forge-Protocol - validation and patch-gating for safer iterative code changes
• bounty-bot - evidence tracking and structured bug bounty workflow
• infj-bot - companion reasoning layer and internal AI workflow surface
• LOTUS-ACADEMY - hands-on security training and operator practice
• cyber-tools - scoped utilities for research and validation

ANCHOR is the cockpit. The other repos are proof that the idea generalizes.

Start here: ARCHITECTURE.md — canonical repo ownership, data flow, and merge policy.

Consolidating GitHub and local trees so only finished, reusable code stays public:

• ROADMAP.md — engineering phases (A–D) and Phase C priorities
• ARCHITECTURE.md — definitive component map (ANCHOR / AFP / infj-bot)
• REPO_CONSOLIDATION_PLAN.md — tiers, archive list, phases
• SCRIPT_REUSE_MAP.md — which scripts move into which repo
• ./scripts/github_repo_audit.sh — refresh audit markdown
• ./scripts/github_repo_cleanup.sh --dry-run — preview fork archives
• ./scripts/verify_duplicate_repo_deps.sh — before deleting local duplicate dirs

• Reproducibility over assertion
• Honest states over polished fiction
• Scope and authorization first
• Evidence does the heavy lifting
• Smaller claims, stronger proof

• structured reasoning systems
• reproducible validation
• evidence-first workflows
• security research tooling
• local-first systems that people can actually run

SHA256:CwsTnS0/92Ymo/TXGQ+iRVfILLN2Q2GZBpN/WFLxOMI

• Immunefi: timeless-hayoka
• X: @timelesshayoka
• Portfolio: web3-security-portfolio