{
  "markdown": "<!-- mcp-name: io.github.gowtham0992/redline -->\n\n<h1 align=\"center\">redline</h1>\n\n<h2 align=\"center\">Catch prompt regressions before they ship.</h2>\n\n<p align=\"center\">\n  <strong>Automatic eval suites from the prompt logs you already have.</strong>\n</p>\n\n<p align=\"center\">\n  redline turns real prompt-response logs into local regression tests. It selects\n  representative cases, replays your changed prompt, and shows the behavioral\n  diff before a bad prompt reaches users.\n</p>\n\n<p align=\"center\">\n  <a href=\"https://gowtham0992.github.io/redline/\">Website</a> ·\n  <a href=\"#project-docs\">Docs</a> ·\n  <a href=\"docs/mcp.md\">MCP</a> ·\n  <a href=\"https://registry.modelcontextprotocol.io/?q=io.github.gowtham0992%2Fredline\">MCP Registry</a> ·\n  <a href=\"SECURITY.md\">Security</a> ·\n  <a href=\"LICENSE\">License</a>\n</p>\n\n[![CI](https://github.com/gowtham0992/redline/actions/workflows/ci.yml/badge.svg?branch=main)](https://github.com/gowtham0992/redline/actions/workflows/ci.yml)\n[![GitHub Pages](https://github.com/gowtham0992/redline/actions/workflows/pages.yml/badge.svg?branch=main)](https://github.com/gowtham0992/redline/actions/workflows/pages.yml)\n[![PyPI](https://img.shields.io/pypi/v/redline-ai.svg)](https://pypi.org/project/redline-ai/)\n[![MCP Registry](https://img.shields.io/badge/MCP%20Registry-io.github.gowtham0992%2Fredline-blue)](https://registry.modelcontextprotocol.io/?q=io.github.gowtham0992%2Fredline)\n[![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)\n[![Stars](https://img.shields.io/github/stars/gowtham0992/redline?style=social)](https://github.com/gowtham0992/redline/stargazers)\n\n## Start Here\n\nInstall from PyPI:\n\n```bash\npython -m pip install redline-ai\n```\n\nRun the guided local app with the public proof loaded:\n\n```bash\nredline app --demo\n```\n\nThis generates the public demo reports, opens the local product app, and shows\nthe full import -> suite -> eval -> review workflow. The demo catches ten\nsynthetic regressions without API keys, private logs, a cloud account, or an LLM\njudge.\n\nPrefer terminal output first:\n\n```bash\nredline demo --public --compact\n```\n\nThe demo writes JSON, Markdown, and self-contained HTML reports under\n`.redline/demo`.\n\nAsk redline what to do next:\n\n```bash\nredline status --reports-dir .redline/demo/reports\n```\n\n`status` reads local config, suites, reports, history, and audit evidence, then\nprints the next command instead of leaving you to infer the workflow.\n\nOpen the guided local product app on existing reports:\n\n```bash\nredline app --reports-dir .redline/demo/reports\n```\n\nThe app is a local, copy-command workflow: import logs, generate suites, run\nevals, review regressions, record history, and export CI/MCP setup without the\nbrowser executing shell commands.\n\nOn headless CI or remote shells, skip browser opening and use the printed HTML\npath:\n\n```bash\nredline app --reports-dir .redline/demo/reports --no-open --out .redline/app.html\n```\n\n<details>\n<summary>First-run troubleshooting</summary>\n\n- `redline: command not found`: run `python -m pip install redline-ai`, then\n  confirm `python -m pip show redline-ai`.\n- App did not open: use `--no-open --out .redline/app.html` and open or\n  upload that file from your environment.\n- Suite not found: run `redline suite logs/baseline.jsonl --out redline-suite.json`.\n- Validation failed: run `redline validate redline-suite.json --strict` and fix\n  the first reported error.\n- GitHub Action cannot find a suite: commit `redline-suite.json` or point the\n  action `suite` input at your prompt manifest.\n\nFull guide: [docs/troubleshooting.md](docs/troubleshooting.md).\n</details>\n\n![redline product demo](https://raw.githubusercontent.com/gowtham0992/redline/main/site/assets/redline-product-demo.gif)\n\n## Product Proof\n\nredline has two proof paths: a fast first-run demo and a larger public-data\ndogfood run.\n\n| Proof | Command or data | Result |\n| --- | --- | --- |\n| First-run demo | `redline demo --public --compact` | 10 synthetic regressions caught locally with no API keys. |\n| Internet dogfood | 100 prompt-response rows sampled from [Databricks Dolly 15k](https://huggingface.co/datasets/databricks/databricks-dolly-15k) | 51 regressions, 27 changed cases, 22 neutral controls, and 0 dashboard warnings. |\n| Release gate | tests, lint, type check, action smoke, and release build | Package, CI, report, dashboard, and MCP paths are validated before publish. |\n\nThese screenshots are local artifacts from the 100-row internet dogfood run.\n\n| Dashboard | HTML report |\n| --- | --- |\n| ![redline dashboard showing reports, benchmark evidence, history, and ship readiness](https://raw.githubusercontent.com/gowtham0992/redline/main/site/assets/redline-dashboard-proof.png) | ![redline HTML report showing concrete regression reasons and side-by-side baseline and candidate outputs](https://raw.githubusercontent.com/gowtham0992/redline/main/site/assets/redline-report-proof.png) |\n\n## What Is redline?\n\nredline is an open-source, local-first eval tool for AI teams. It uses logs you\nalready have: prompts, outputs, support tickets, traces, model responses, and\nproduction JSONL exports.\n\nInstead of asking you to hand-write evals first, redline generates the first\nsuite from real behavior. You can then run that suite every time a prompt,\nmodel, or runner changes.\n\nNo cloud account is required. No manual test writing is required. No LLM judge\nis required for the core regression signal. The package has zero runtime\ndependencies, which keeps installs fast and the default supply-chain surface\nsmall.\n\n## How It Works\n\nredline gives you three primitives that cover the prompt-regression loop:\n\nFor a first pass on two local logs, use one command:\n\n```bash\nredline quick-check logs/baseline.jsonl logs/candidate.jsonl --open\n```\n\nIt generates a temporary suite, writes JSON/Markdown/HTML reports plus a guided\nlocal app under `.redline/quick-check`, opens the focused HTML report, and\nprints the concrete behavioral diff. Use `--open-app` when you want the guided\nreview workflow to open instead of only the focused report.\n\n### 1. Logs\n\nStart with prompt-response data you already have. Import JSONL, convert exports\nfrom tools like Langfuse or Helicone, capture OpenAI/Anthropic SDK calls, or add\nbounded FastAPI/ASGI middleware.\n\n```bash\nredline import downloaded.jsonl --detect\nredline import downloaded.jsonl --auto-map --preview 3\nredline import downloaded.jsonl --auto-map --out logs/baseline.jsonl\nredline import downloaded.jsonl --input-field instruction --output-field response --preview 3\nredline import downloaded.jsonl --input-field instruction --output-field response --out logs/baseline.jsonl\nredline import langfuse-export.jsonl --preset langfuse --out logs/baseline.jsonl\nredline suite logs/baseline.jsonl --out redline-suite.json\nredline cases redline-suite.json\n```\n\nUse `--detect` when you do not know the field names. Use `--preview` when the\nexport is new to you; it shows mapped, redacted sample rows without writing a\nbaseline file.\n\nSuite generation prints a readiness score and improvement suggestions. That\nscore measures suite health, not model quality or candidate safety.\n\n### 2. Suite\n\nredline groups behavior into deterministic signatures and selects\nrepresentative cases first. You can add pinned edge cases and explicit\nrequirements when a scenario must never be missed.\n\n```bash\nredline cases redline-suite.json\nredline suite add redline-suite.json --prompt \"...\" --response \"...\"\n```\n\n### 3. Eval\n\nReplay a changed prompt or compare candidate outputs. redline names the\nbehavior that broke: missing JSON keys, URLs, numbers, tables, code blocks,\nrefusals, empty answers, or requirement failures.\n\n```bash\nredline eval --prompt prompts/v2.txt\nredline diff redline-suite.json logs/candidate.jsonl\n```\n\n## Product Promise\n\nIn under five minutes, on a real prompt log, redline should catch one regression\nyou did not want to ship.\n\nThat promise is intentionally narrow. redline is not a hosted eval platform, a\ngeneric score, or a replacement for human judgment. It is the local safety loop\nbetween \"I changed the prompt\" and \"this is safe enough to merge.\"\n\n## Real Workflow\n\nBuild a suite from baseline logs:\n\n```bash\nredline suite logs/baseline.jsonl --out redline-suite.json\n```\n\nEvaluate a changed prompt file through your configured runner:\n\n```bash\nredline eval --prompt prompts/v2.txt\n```\n\nOr compare candidate outputs you already generated:\n\n```bash\nredline diff redline-suite.json logs/candidate.jsonl\n```\n\nWhen redline finds a blocking change, it exits non-zero for CI and prints the\nreason:\n\n```text\nREGRESSION case_004\n- candidate missing JSON keys: owner, required_action\n- candidate missing URL: https://example.com/policies/refunds\n\nConfidence: HIGH | fix blocking cases before shipping\n```\n\n## What redline Catches\n\n| Signal | Example regression |\n| --- | --- |\n| JSON validity and keys | Candidate stops returning valid JSON or drops `owner`. |\n| Tables, lists, and code blocks | Markdown table becomes prose; code fence disappears. |\n| Numbers, URLs, and entities | Refund window, ticket ID, policy URL, or owner is missing. |\n| Empty outputs and refusals | Candidate newly refuses a safe task or returns nothing. |\n| Content drift | Same-shape response changes substantially. |\n| Explicit requirements | Pinned cases require or forbid exact strings. |\n\nredline is deterministic and local-first by default. Optional judge commands are\navailable for ambiguous `changed` cases, but redline does not call a cloud model\nunless you explicitly configure that command.\n\nThat is the point. redline is designed to be the fast merge-blocking gate for\nregressions that break production systems: invalid JSON, missing required\nfields, lost tables, empty answers, dropped URLs, changed refusal behavior, and\nexplicit requirement failures. LLM judges are useful for semantic review, but\nthey are slower, cost money, and can be flaky in CI. redline keeps the default\ngate deterministic, reproducible, and cheap, then lets you add judges only where\nthe structural signal is not enough.\n\nMethodology details live in [docs/methodology.md](docs/methodology.md).\n\nSuite generation does not run statistical or embedding clustering by default.\nIt groups logs by deterministic behavior signatures, such as prompt intent,\nresponse shape, length bucket, and JSON schema. It picks one representative per\ngroup first, then adds high-variance edges and evenly spread prompt-diverse\nsamples from large groups when the case budget allows.\n\n## Trust Boundary\n\nA green redline run means no configured high-signal structural blockers were\nfound. It does not prove factual correctness, tone, hallucination safety, policy\ncompliance, or subtle reasoning quality.\n\nThat boundary is visible in CLI output and reports because over-trusting eval\ntools is dangerous. Each reported case includes a confidence and signal\n(`structural`, `shallow_semantic`, `requirement`, `judge`, or `human_judgment`)\nso reviewers can see why redline is making the call. Use requirements or an\noptional judge for semantic risks that structural checks cannot prove.\n\n## Product Surface\n\nredline is built around the full prompt-regression loop:\n\n- `redline watch`: collect prompt-response observations from logs, Python\n  functions, OpenAI/Anthropic-compatible SDK calls, or ASGI apps, with\n  best-effort common secrets and PII redacted before write by default.\n- `redline import`: normalize exported team logs into redline JSONL, with the\n  same best-effort redaction enabled by default. Use `--no-redact` only for\n  reviewed local-only logs.\n- `RedlineMiddleware`: capture bounded JSON FastAPI or ASGI request/response pairs locally, with optional skip diagnostics.\n- `redline redact --check`: scan logs for common secrets and PII, then write a scrubbed copy when needed.\n  Redaction is best-effort pattern matching, not a privacy boundary; review sensitive logs before sharing.\n- `redline cluster`: inspect deterministic behavior-signature groups before suite generation.\n- `redline suite`: generate a representative eval suite from baseline logs.\n- `redline prompts`: scan many prompt files and write or check a versionable prompt-to-suite manifest.\n  Add `--check-suites` in CI when every prompt should already have a built and valid suite.\n- `redline suite add`: pin hand-picked edge cases the algorithm should never miss.\n- `redline budget` / `redline benchmark`: estimate suite or prompt-manifest runtime without\n  executing replay commands, write budget artifacts, and optionally fail on a CI\n  time budget. Add `--measure-local` to time redline's deterministic local diff\n  work on your suite baselines without calling a model.\n- `redline eval`: replay each suite case through your local app or model runner.\n- `redline diff`: compare candidate JSONL outputs against the suite baseline.\n- `redline mark` and `redline accept`: review intentional changes and promote the\n  new baseline.\n- `redline require`: add deterministic must-include or must-not-include rules.\n- `redline audit --verify`: inspect the local audit trail and verify the hash chain.\n  Add `--expect-last-hash` or `--expect-entries` when you want to prove the\n  local log tail still matches a checkpoint from CI or release evidence. Add\n  `--out-checkpoint .redline/audit-checkpoint.json` to persist that evidence,\n  then `--checkpoint .redline/audit-checkpoint.json` to verify against it later.\n- `redline sbom`: write CycloneDX SBOM release evidence for security review.\n- `redline app`: open the guided local product surface for importing logs,\n  generating suites, reviewing regressions, recording history, and wiring CI/MCP.\n- `redline status`: show project readiness and the next command from local\n  evidence, including the guided app command, first review case, its reason, and\n  why it matters.\n- `redline history`, `redline compare`, and `redline dashboard`: track quality\n  over time and inspect report artifacts locally. The dashboard surfaces\n  feature-level rollups, prompt-level eval rows, benchmark evidence, and a\n  latest-report review queue when reports come from a prompt manifest. It also\n  warns when reports exist without benchmark evidence from the same project.\n- `redline summary`: inspect suite readiness, or pass `redline-prompts.json` to\n  roll up multi-prompt suite coverage, owners, requirements, and missing suites.\n- `redline-mcp`: let AI coding assistants run checks inside Claude, Codex,\n  Cursor, Kiro, or any MCP client.\n\nFor repos with many prompt files, the manifest becomes the eval plan:\n\n```bash\nredline prompts prompts/ --suite-dir suites --out redline-prompts.json\nredline prompts prompts/ --suite-dir suites --out redline-prompts.json --check --check-suites\nredline summary redline-prompts.json\nredline validate redline-prompts.json --strict\nredline budget redline-prompts.json\nredline eval redline-prompts.json\n```\n\nManifest summaries show readiness across every mapped suite, manifest\nvalidation checks every mapped suite, manifest benchmarks aggregate runtime\nbudget, and manifest evals print prompt-level rollups before case details. Large\nrepos can see which prompt files or feature folders need attention first.\n\nWhen mapped suites are valid, the check prints ready commands such as:\n\n```bash\nredline eval suites/support/triage.redline-suite.json --prompt prompts/support/triage.txt\n```\n\n## Connect Your App\n\nAny command that reads a prompt from stdin and prints a response to stdout can\nbe a redline runner:\n\n```bash\nredline init --runner stdio --copy-runner --github-action\n```\n\nBuilt-in adapters cover provider-neutral stdio, OpenAI, Anthropic, LiteLLM,\nHTTP APIs, Python chains, JSONL log imports, and OpenAI/Anthropic SDK capture:\n\n```bash\nredline runners\nredline runners --copy all\n```\n\nRunner details live in [docs/runners.md](docs/runners.md). Log import and SDK\ncapture adapters are for building suites from real observations, not for\n`redline eval` replay. The JSONL log adapter includes Langfuse, Helicone,\nLangSmith, and Braintrust presets for exported observability logs.\n\n## AI Assistant Native\n\nredline ships a local Model Context Protocol server:\n\n```bash\nredline-mcp\n```\n\nUse [docs/mcp.md](docs/mcp.md) to wire redline into an MCP client. The MCP\nsurface exposes safe capture-readiness, privacy, audit, scale, read,\nquick-check, case-inspection, eval, and report tools plus workflow prompts like\n`setup_redline_project`, `check_prompt_change`, `build_suite_from_logs`, and\n`review_candidate_outputs`.\nIt can also list or copy runner adapters and optional judge templates during setup.\nThe only mutating MCP tool is guarded: `redline_mark` requires `allow_write: true`\nand a note before it records an intentional case judgment. Baseline promotion\nstays CLI-only.\n\n## CI And GitHub\n\nCreate config plus a GitHub Actions workflow:\n\n```bash\nredline init --runner stdio --copy-runner --github-action\n```\n\nUse redline as a composite GitHub Action from another repo:\n\n```yaml\n- uses: gowtham0992/redline@v0.3.0\n  with:\n    prompt-path: prompts/v2.txt\n    benchmark-max-seconds: \"300\"\n```\n\nFor multi-prompt repos, point `suite` at `redline-prompts.json`. The action\nchecks every mapped suite with `redline prompts --check --check-suites`, runs a\nmanifest-wide benchmark, then runs the manifest eval.\n\nThe action writes JSON, full Markdown, concise PR-comment Markdown, HTML, JUnit,\nSlack-ready JSON, history, dashboard, and audit checkpoint artifacts under\n`.redline/`, appends benchmark, concise eval, and trend summaries to the GitHub\nstep summary, and exits with the eval gate status. Set `benchmark-max-seconds`\nwhen a suite should fail CI if its worst-case runtime budget grows too far.\n\n## Reports\n\nEvery `diff` and `eval` run can write:\n\n- JSON for machines and dashboards\n- full Markdown for detailed summaries, including prompt-manifest rollups\n- concise PR-comment Markdown for merge-review surfaces\n- self-contained HTML for side-by-side inspection, including feature and prompt eval tables\n- JUnit XML for CI test reporting\n- Slack Block Kit JSON for CI bots or webhook integrations you control\n- GitHub annotations for changed or blocking cases\n\nExample:\n\n```bash\nredline diff redline-suite.json logs/candidate.jsonl \\\n  --out-json .redline/reports/diff.json \\\n  --out-md .redline/reports/diff.md \\\n  --out-comment .redline/reports/diff-comment.md \\\n  --out-html .redline/reports/diff.html \\\n  --out-junit .redline/reports/diff.xml \\\n  --out-slack .redline/reports/diff.slack.json\n```\n\n## Optional Judges\n\nUse judges only where structural checks are not enough. redline sends only\nambiguous `changed` cases to the configured command as JSON on stdin:\n\n```bash\nredline judges\nredline judges --copy openai\nredline judges --copy support-rubric\nredline diff logs/candidate.jsonl --judge \"python examples/judge_changed.py\"\n```\n\nRepo examples and installable templates:\n\n- [examples/judge_changed.py](examples/judge_changed.py)\n- [examples/openai_judge.sh](examples/openai_judge.sh)\n- [examples/anthropic_judge.sh](examples/anthropic_judge.sh)\n- [examples/litellm_judge.sh](examples/litellm_judge.sh)\n- [examples/judges/support_rubric.md](examples/judges/support_rubric.md)\n- [examples/judges/extraction_rubric.md](examples/judges/extraction_rubric.md)\n- [examples/judges/safety_rubric.md](examples/judges/safety_rubric.md)\n\nCalibration guidance lives in [docs/judges.md](docs/judges.md).\n\n## Config\n\n`redline init` writes `redline.json` with a `$schema` reference for editor\nautocomplete. Important keys:\n\n| Key | Purpose |\n| --- | --- |\n| `suite` | Suite baseline path, default `redline-suite.json`. |\n| `input_field`, `output_field` | JSONL field paths for prompts and responses. |\n| `max_cases` | Maximum representative cases selected for a suite. |\n| `replay` | Command used by `eval`; prompts go to stdin by default. `{prompt}` is for small legacy argv runners; `{prompt_file}` passes a temporary rendered-prompt file path. |\n| `workers` | Number of replay cases to run concurrently. |\n| `owners` | Optional pattern-to-owner rules so regressions show the responsible team. |\n| `approval` | Optional local guardrail; `require_approver` makes `accept` record an approver. |\n| `fail_on` | Statuses that fail `diff` or `eval`; use `\"none\"` for report-only setup. |\n| `reports` | JSON, Markdown, PR-comment Markdown, HTML, JUnit, and Slack-ready JSON output paths. |\n| `logs` | Observed prompt-response log path and optional middleware skip diagnostics path. |\n| `audit` | Append-only JSONL audit log path for evals, judgments, requirements, and accepted baselines. New entries include operator/approver context plus a local hash chain that `redline audit --verify` can check; use expected hash/count checkpoints or `--out-checkpoint` evidence files to detect tail truncation. |\n| `judge` | Optional command for ambiguous `changed` cases. |\n\nCheck setup before relying on a suite:\n\n```bash\nredline doctor --strict\nredline validate redline-suite.json --strict\nredline summary redline-suite.json\n```\n\n`doctor` shows whether the suite has explicit requirements or recorded\njudgments before you rely on structural checks in CI.\n`summary` reports a suite readiness score, behavior-group/case coverage, owner\ncoverage, accepted baseline history, approver coverage, and explicit guard\ncoverage for cases with requirements or recorded judgments so teams can review\nsuite readiness before CI.\n`dashboard` also shows audit checkpoint evidence when `.redline/audit-checkpoint.json`\nis present.\n\n## Dogfood Assets\n\nThe public fixture is synthetic, shaped after public instruction/chat dataset\npatterns, and documented in\n[examples/public_dogfood_sources.md](examples/public_dogfood_sources.md).\n\n```bash\npython -m redline suite examples/public_dogfood_baseline.jsonl --out /tmp/redline-public-suite.json --all-cases\npython -m redline diff /tmp/redline-public-suite.json examples/public_dogfood_candidate.jsonl --compact --fail-on none\n```\n\nFor AI-assistant session dogfood, use\n[docs/ai-session-dogfood-prompts.jsonl](docs/ai-session-dogfood-prompts.jsonl)\nand normalize raw exports with `scripts/normalize_ai_session_logs.py`.\nReproducible dogfood case studies live in\n[docs/case-studies.md](docs/case-studies.md).\nPublic dataset candidates for internet dogfood are ranked in\n[docs/internet-dogfood-sources.md](docs/internet-dogfood-sources.md).\n\nFrom a repo checkout, record the public demo:\n\n```bash\nbash scripts/demo_terminal.sh\nbash scripts/demo_gif.sh .redline/launch .redline/launch/redline-demo.gif\n```\n\n## Development\n\n```bash\npython -m pip install -e \".[dev]\"\npython -m pytest -q\npython -m ruff check .\npython -m mypy redline tests scripts examples\n```\n\nBefore cutting a release or asking someone else to try a branch:\n\n```bash\nbash scripts/release_check.sh\n```\n\n## Project Docs\n\n- [docs/release.md](docs/release.md): package, tag, PyPI, and MCP Registry release flow\n- [docs/launch.md](docs/launch.md): public alpha launch plan\n- [docs/troubleshooting.md](docs/troubleshooting.md): first-run and CI failure recovery\n- [docs/import-guides.md](docs/import-guides.md): Langfuse, Helicone, OpenAI chat, Datadog, and custom log import recipes\n- [docs/methodology.md](docs/methodology.md): behavior grouping, case selection, scoring, and trust boundaries\n- [docs/calibration.md](docs/calibration.md): tiny fixture showing regressions, changed cases, and neutral cases\n- [docs/commands.md](docs/commands.md): compact CLI command reference\n- [docs/real-log-dogfood.md](docs/real-log-dogfood.md): redaction-first real-log test protocol\n- [docs/dogfood.md](docs/dogfood.md): first-user dogfood protocol\n- [docs/case-studies.md](docs/case-studies.md): reproducible dogfood case studies\n- [docs/internet-dogfood-sources.md](docs/internet-dogfood-sources.md): public prompt-response datasets for dogfood sourcing\n- [docs/runners.md](docs/runners.md): runner and log adapter setup\n- [docs/mcp.md](docs/mcp.md): MCP server setup\n- [docs/benchmarks.md](docs/benchmarks.md): performance contract and CI benchmark artifacts\n- [docs/repository.md](docs/repository.md): GitHub repository controls\n- [scripts/README.md](scripts/README.md): maintainer script index\n- [CONTRIBUTING.md](CONTRIBUTING.md): contributor validation\n- [SECURITY.md](SECURITY.md): privacy and vulnerability reporting\n- [LICENSE](LICENSE): MIT open source license\n\nWebsite source for GitHub Pages lives in [site/](site/) and deploys from the\ncommitted static assets on `main`.\n",
  "bytes": 24421,
  "sha": "3fbc8690348342f318e887475c1f0d5e85878bc5b1f3982e86ad8de5d055bd92",
  "repo_slug": "gowtham0992/redline",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_gowtham0992_redline_2df580a9/readme"
}