[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-92318":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":15,"stars30d":16,"stars90d":15,"forks30d":15,"starsTrendScore":15,"compositeScore":17,"rankGlobal":9,"rankLanguage":9,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":19,"topics":22,"createdAt":9,"pushedAt":9,"updatedAt":23,"readmeContent":24,"aiSummary":25,"trendingCount":15,"starSnapshotCount":15,"syncStatus":26,"lastSyncTime":27,"discoverSource":28},92318,"Autonomous-Forge","OmarH-creator\u002FAutonomous-Forge","OmarH-creator","This repo is automated and is fully AI-built and AI-maintained project that continuously plans, codes, tests, and improves itself.",null,"Python",247,9,13,1,0,167,53,"Apache License 2.0",false,"main",true,[],"2026-07-22 04:02:05","# Autonomous Forge\n\nAutonomous Forge is an open-source, AI-built and AI-maintained developer tool for safely running repository-native autonomous software-improvement loops.\n\nThe project starts as a local-first Python CLI. Its first goal is deliberately small: provide a `forge` command that can grow into dry-run planning, task selection, validation reporting, diff review, supplied commit-status review, combined change readiness, guarded patch preview, explicitly confirmed patch application, post-apply validation handoff, commit-readiness review, commit metadata preview, explicitly confirmed local commit creation, post-commit verification, commit trust review, branch-protection-aware push-readiness review, an explicitly confirmed fast-forward-only non-force push handoff, post-push verification, hash-linked durable maintenance evidence bundles, persisted bundle verification, replay summaries with compact replay policy gates, run-history bundle links, run-history link quality review, linked-bundle replay verification from history pointers, reviewer-facing preservation handoffs with history\u002Fbundle context consistency, comparison-oriented maintenance handoff summaries, and durable repository memory without requiring uncontrolled autonomous behavior.\n\nFor a visual orientation to the current workflow and its safety boundary, see [the project overview](docs\u002FOVERVIEW.md).\n\n## Current Autonomous Status\n\nAutonomous Forge is pre-alpha. Latest autonomous run: AUTO-126 tightened `forge maintenance-review-handoff` so a ready reviewer handoff now requires the `.ai\u002Frun-history` pointer and replayed linked bundle to agree on reviewed paths, validation steps, and retained validation context. This blocks stale or manually edited history links that still reference a hash-valid bundle but no longer describe the same reviewed maintenance change. Validation was limited to static source\u002Ftest\u002Fdocs review through the GitHub repository API plus scratch syntax compilation of the new helper logic because full repository checkout\u002Ffull pytest is unavailable in this runtime. No visual updates were needed because the existing overview remains accurate. Current limitations: handoffs still rely on persisted JSON evidence and recomputed hashes; they do not rerun validation, poll workflows, prove signature identity, or prove validation coverage. Next autonomous objective: carry the new context-consistency gate into maintenance review comparisons so multi-run summaries expose which completed records have pointer\u002Fbundle drift.\n\nThe repository now contains:\n\n- Apache-2.0 licensing and durable planning files in `.ai\u002F`.\n- A minimal Python package with a primary `forge` console script plus compatibility scripts for the safer maintenance workflow commands.\n- Task parsing, deterministic task selection, roadmap linting, repository reports, policy summaries, repository inventory, enriched implementation plans, plan-enriched change proposals, enriched validation plans, validation-run previews, validation orchestration previews, enriched command-execution handoff previews, enriched executor precondition gates, enriched executor contract previews, enriched executor dry-run previews, one narrow opt-in executor run command with explicit result-persistence handoff, changed-file\u002Fcontent review, git-diff review, commit\u002Fworkflow status review, guarded patch preview\u002Fapply, post-apply validation, commit readiness\u002Fcreate\u002Fverify\u002Ftrust review, branch-protection-aware push readiness, branch-policy-enforcing push handoff, post-push verification, durable evidence bundles, persisted bundle verification\u002Freplay summaries, run-history links for completed pushed maintenance bundles, run-history link quality review with strict linked-bundle replay verification, reviewer-facing maintenance handoffs with history\u002Fbundle context consistency, comparison-oriented maintenance handoff summaries, validation-result writes that retain implementation context in saved records, run-history read\u002Fcompare surfaces that expose retained validation context, replay summaries that expose retained validation context when persisted bundles include it, maintenance bundle creation\u002Fhistory links that preserve validation context automatically, replay-summary consistency checks that compare retained context with reviewed paths and preserved validation steps, and replay policy-gate summaries for compact pass\u002Ffail\u002Fadvisory replay review.\n- Deterministic tests for the CLI’s current local workflows, including enriched `forge plan`, `forge propose`, `forge validate-plan`, `forge validation-preview`, `forge validation-orchestration`, `forge command-execution-handoff`, `forge executor-gate`, `forge executor-contract`, `forge executor-dry-run`, `forge executor-run`, context-retaining validation-result writes, validation-context-aware run-history reads\u002Fcomparisons, validation-context-aware maintenance replay summaries, validation-context-preserving maintenance bundle\u002Fhistory links, history-link quality review and strict linked-bundle replay verification, reviewer-facing maintenance handoffs, handoff history\u002Fbundle context consistency, maintenance review comparison summaries, replay validation-context consistency checks, replay policy-gate summaries, branch-policy push handoff, allowed-signer commit trust, branch-protection-aware push readiness, and maintenance bundle history-link behavior.\n- CI smoke coverage that validates the live roadmap, installed console entry points, extension help routes, and the test suite across Python 3.10, 3.11, and 3.12.\n\n## Install for local development\n\n```bash\npython -m pip install -e .\nforge --help\nforge plan --help\nforge propose --help\nforge validate-plan --help\nforge git-diff-review --help\nforge patch-apply --help\nforge commit-create --help\nforge commit-verify --help\nforge push-readiness --help\nforge push-handoff --help\nforge maintenance-evidence-bundle --help\nforge maintenance-history-link-review --help\nforge maintenance-review-handoff --help\nforge maintenance-review-compare --help\nforge-maintenance-replay-policy-summary --help\n```\n\nFor full setup, contribution workflow, and safety expectations, see `CONTRIBUTING.md`.\n\n## Core planning and review workflow\n\n```bash\nforge tasks --plan .ai\u002FAUTONOMOUS_PLAN.md --next\nforge plan --plan .ai\u002FAUTONOMOUS_PLAN.md --state .ai\u002FAUTONOMOUS_STATE.md --policy .forge\u002Fpolicy.md --root .\nforge propose --plan .ai\u002FAUTONOMOUS_PLAN.md --state .ai\u002FAUTONOMOUS_STATE.md --policy .forge\u002Fpolicy.md --root .\nforge validate-plan --plan .ai\u002FAUTONOMOUS_PLAN.md --state .ai\u002FAUTONOMOUS_STATE.md --policy .forge\u002Fpolicy.md --root .\nforge validation-orchestration --plan .ai\u002FAUTONOMOUS_PLAN.md --state .ai\u002FAUTONOMOUS_STATE.md --policy .forge\u002Fpolicy.md --root .\nforge executor-dry-run --plan .ai\u002FAUTONOMOUS_PLAN.md --state .ai\u002FAUTONOMOUS_STATE.md --policy .forge\u002Fpolicy.md --root . --command \"python -m pytest\" --confirm-executor-dry-run\nforge executor-run --plan .ai\u002FAUTONOMOUS_PLAN.md --state .ai\u002FAUTONOMOUS_STATE.md --policy .forge\u002Fpolicy.md --root . --command \"python -m pytest\" --confirm-executor-dry-run\nforge maintenance-history-link-review --link .ai\u002Frun-history\u002FAUTO-120-link.json --require-linked-replayable\nforge maintenance-review-handoff --link .ai\u002Frun-history\u002FAUTO-120-link.json --require-ready\nforge maintenance-review-compare --link .ai\u002Frun-history\u002FAUTO-120-link.json --link .ai\u002Frun-history\u002FAUTO-121-link.json --require-all-ready\n```\n\nMost commands are local-first and read-only unless their contract explicitly requires a confirmation flag for a narrow local write, local validation execution, local commit creation, or non-force push handoff. `forge plan` is read-only: it selects the highest-priority eligible roadmap task, lists policy boundaries, turns roadmap prose into reviewable implementation steps\u002Ffile targets\u002Fvalidation steps\u002Frisks, and never changes repository state. `forge propose`, `forge validate-plan`, `forge validation-preview`, `forge validation-orchestration`, `forge command-execution-handoff`, `forge executor-gate`, `forge executor-contract`, `forge executor-dry-run`, and `forge executor-run` carry those same structured fields forward into downstream handoff artifacts. `forge executor-run` may run one exact confirmed validation command with `shell=false`, but it only prepares the result-persistence handoff; saved history still requires a separate explicit write confirmation. `forge validation-result-write` retains any existing implementation context from that saved record under `record.validation_context` while attaching the supplied validation result, `forge run-history-read` \u002F `forge run-history-compare` expose that retained context for later audit, and `forge maintenance-evidence-bundle` preserves retained validation context in newly generated bundles\u002Fhistory links so `forge maintenance-history-link-review` can first review run-history pointer quality and, with `--require-linked-replayable`, verify the linked bundle SHA-256 and run `forge maintenance-replay-summary` for hash-linked source-report replay, retained validation-context consistency, and compact replay policy gates. `forge maintenance-review-handoff` packages those results into one reviewer-facing preservation handoff and now requires the history pointer's reviewed paths, validation steps, and retained validation context to match the replayed linked bundle before reporting ready. `forge maintenance-review-compare` compares multiple handoffs so a reviewer can see which completed evidence records are ready or blocked without opening each raw bundle. `forge-maintenance-replay-policy-summary` exposes the same compact gates as a dedicated compatibility command for bundle-first review workflows.\n\n## Run tests\n\n```bash\npython -m pip install -e . pytest==8.3.5\npython -m pytest -q\n```\n\n## Safe contribution expectations\n\nContributions should stay small, local-first, and reviewable. Do not add network actions, external command execution, secret handling, deployment behavior, telemetry, or repository-permission changes unless the roadmap and repository policy explicitly allow it.\n","Autonomous Forge 是一个开源的、由 AI 构建并自主维护的开发者工具，旨在安全地在代码仓库本地运行闭环式的软件自改进流程。它以 Python CLI 为核心，提供逐步受控的自动化能力，包括任务规划、变更提案、验证预览、人工确认式补丁应用、提交与推送决策、历史回放验证及持久化维护证据链等，所有操作均强调可审计性、可追溯性和人工干预边界。项目采用分阶段演进策略，当前处于 pre-alpha 阶段，聚焦于静态审查、语法校验与哈希锚定的历史一致性保障。适用于重视工程可控性与合规性的中大型开源项目或企业级代码库的渐进式自动化维护场景。",2,"2026-07-08 04:30:04","CREATED_QUERY"]