RepoRevenant — Forgotten Code Deserves a Second Life

说明与用途

RepoRevenant 通过本地 LLM、确定性 AST 分析和沙箱测试验证，恢复旧版 Python 仓库，实现零云泄露和可审计性。

问题

Abandoned Python repositories often fail on modern runtimes due to removed modules, API drift, and syntax changes, preventing reuse of valuable code.

结果

RepoRevenant successfully resurrects legacy Python repos by combining deterministic AST checks, local LLM repair synthesis, human approval, and sandboxed test verification, achieving zero‑cloud leakage and auditability.

工作流

• Validate HTTPS Git URL with SSRF protection
• Parse requirements and discover test suites
• Run deterministic AST pass to flag PEP 594 and other issues
• Execute baseline pytest in isolated sandbox
• Send AST evidence to local LLMs (llama3:8b, mistral:7b, codellama:7b) for root‑cause hypotheses and repair synthesis
• Validate AI‑generated diffs with ast.parse(), create rollback backup
• Apply patch after human approval
• Run pytest again to confirm exit code 0
• Export Resurrection Certificate, telemetry, and patch files

证据

RepoRevenant inspects repository structure, extracts empirical Abstract Syntax Tree (AST) evidence, isolates the code in clean temporary sandboxes, executes baseline test suites to document exact pre‑patch failures, synthesizes targeted repairs using replaceable open‑weight AI models, and proves the resurrection by verifying post‑patch test suites pass with exit code 0.

来源

The theme "Ghost in the Machine" maps directly: the ghost is the dormant intelligence trapped inside an abandoned codebase.

来源