Skip to content

aiskillstore/marketplace / skills/2025emma/claude-cookbooks

claude-cookbooks

Flagged

Claude AI cookbooks - code examples, tutorials, and best practices for using Claude API. Use when learning Claude API integration, building Claude-powered applications, or exploring Claude capabilities.

FileKindSize
SKILL.mdskill md8.5 KB
references/CONTRIBUTING.mdmarkdown6.1 KB
references/README.mdmarkdown5.7 KB
references/capabilities.mdmarkdown2.9 KB
references/index.mdmarkdown981 B
references/main_readme.mdmarkdown5.7 KB
references/multimodal.mdmarkdown1.8 KB
references/patterns.mdmarkdown708 B
references/third_party.mdmarkdown1.3 KB
references/tool_use.mdmarkdown1.7 KB
scripts/memory_tool.pyscript12.9 KB
skill-report.jsonconfig25.9 KB

skill-report.json

View on GitHub
{
  "schema_version": "2.0",
  "meta": {
    "generated_at": "2026-07-18T09:19:17.821Z",
    "slug": "2025emma-claude-cookbooks",
    "source_url": "https://github.com/2025Emma/vibe-coding-cn/tree/main/i18n/en/skills/claude-cookbooks",
    "source_ref": "main",
    "model": "claude",
    "analysis_version": "3.0.0",
    "source_type": "community",
    "content_hash": "52c3d65a42123cff12bf936bd7ba02e6763a544b4cce1dbb831c17762842fed7",
    "tree_hash": "1f415d1e9c33e03319ff2fdd0e86b4f359ca7aa7e44b5d953ef50d5feee04477",
    "provenance": {
      "model_effective": "claude",
      "fallback_chain": [
        {
          "agent": "codex",
          "outcome": "failed",
          "error": "2026-07-18T09:19:18.072715Z ERROR codex_models_manager::manager: failed to refresh available models: Missing environment variable: `HELM_API_KEY`.\n2026-07-18T09:19:18.093793Z ERROR codex_models_manage…"
        },
        {
          "agent": "claude",
          "outcome": "success"
        }
      ],
      "analysis_version": "3.0.0"
    }
  },
  "skill": {
    "name": "Claude Cookbooks Skill",
    "description": "Educational reference skill with code examples, tutorials, and best practices for building applications with the Claude API, translated from the official Anthropic cookbooks.",
    "summary": "Code examples and guides for building with the Claude API.",
    "icon": "📦",
    "author": "2025Emma",
    "license": "MIT",
    "category": "documentation",
    "tags": [
      "claude-api",
      "documentation",
      "tutorials",
      "code-examples",
      "ai-integration"
    ],
    "supported_tools": [
      "claude",
      "codex",
      "claude-code"
    ],
    "risk_factors": [
      "network",
      "env_access",
      "filesystem"
    ]
  },
  "security_audit": {
    "risk_level": "low",
    "is_blocked": false,
    "safe_to_publish": false,
    "analysis_status": "ok",
    "summary": "This is a benign educational skill translated from the official Anthropic cookbooks. All 19 static findings are false positives: the flagged URLs, environment-variable reads, and shell commands are standard documentation for installing tooling and configuring an API key, and the shutil operations belong to a well-secured reference implementation of Claude's memory tool that validates every path against directory traversal. No malicious intent, data exfiltration, or prompt injection was found.",
    "remediation": [
      {
        "issue": "SKILL.md could not be statically scanned due to invalid frontmatter (a 'TRANSLATED CONTENT:' line precedes the YAML frontmatter delimiter).",
        "suggestion": "Remove the leading 'TRANSLATED CONTENT:' line so the '---' frontmatter block is the first content in SKILL.md, allowing automated scanning.",
        "severity": "low"
      },
      {
        "issue": "The CONTRIBUTING.md documents a pipe-to-shell install command (curl ... | sh) for the uv package manager.",
        "suggestion": "This is upstream documentation copied from Astral. Consider noting that users should review install scripts before piping to a shell, or link to the package manager's release page instead.",
        "severity": "low"
      }
    ],
    "risk_factor_evidence": [
      {
        "factor": "network",
        "evidence": [
          {
            "file": "references/CONTRIBUTING.md",
            "line_start": 17,
            "line_end": 17
          },
          {
            "file": "references/CONTRIBUTING.md",
            "line_start": 27,
            "line_end": 27
          }
        ]
      },
      {
        "factor": "env_access",
        "evidence": [
          {
            "file": "references/CONTRIBUTING.md",
            "line_start": 117,
            "line_end": 117
          },
          {
            "file": "references/CONTRIBUTING.md",
            "line_start": 117,
            "line_end": 117
          }
        ]
      },
      {
        "factor": "filesystem",
        "evidence": [
          {
            "file": "scripts/memory_tool.py",
            "line_start": 307,
            "line_end": 307
          },
          {
            "file": "scripts/memory_tool.py",
            "line_start": 359,
            "line_end": 359
          }
        ]
      }
    ],
    "static_findings": [
      {
        "id": "network:references/CONTRIBUTING.md:17:hardcoded-url",
        "category": "network",
        "severity": "low",
        "pattern": "Hardcoded URL",
        "file": "references/CONTRIBUTING.md",
        "line_start": 17,
        "line_end": 17,
        "snippet": "curl -LsSf https://astral.sh/uv/install.sh | sh"
      },
      {
        "id": "network:references/CONTRIBUTING.md:27:hardcoded-url",
        "category": "network",
        "severity": "low",
        "pattern": "Hardcoded URL",
        "file": "references/CONTRIBUTING.md",
        "line_start": 27,
        "line_end": 27,
        "snippet": "git clone https://github.com/anthropics/anthropic-cookbook.git"
      },
      {
        "id": "env_access:references/CONTRIBUTING.md:117:python-environment-access",
        "category": "env_access",
        "severity": "low",
        "pattern": "Python environment access",
        "file": "references/CONTRIBUTING.md",
        "line_start": 117,
        "line_end": 117,
        "snippet": "api_key = os.environ.get(\"ANTHROPIC_API_KEY\")"
      },
      {
        "id": "env_access:references/CONTRIBUTING.md:117:generic-api-secret-keys",
        "category": "env_access",
        "severity": "high",
        "pattern": "Generic API/secret keys",
        "file": "references/CONTRIBUTING.md",
        "line_start": 117,
        "line_end": 117,
        "snippet": "api_key = os.environ.get(\"ANTHROPIC_API_KEY\")"
      },
      {
        "id": "sensitive:references/CONTRIBUTING.md:48:environment-file-access",
        "category": "sensitive",
        "severity": "high",
        "pattern": "Environment file access",
        "file": "references/CONTRIBUTING.md",
        "line_start": 48,
        "line_end": 48,
        "snippet": "cp .env.example .env"
      },
      {
        "id": "sensitive:references/CONTRIBUTING.md:49:environment-file-access",
        "category": "sensitive",
        "severity": "high",
        "pattern": "Environment file access",
        "file": "references/CONTRIBUTING.md",
        "line_start": 49,
        "line_end": 49,
        "snippet": "# Edit .env and add your Claude API key"
      },
      {
        "id": "blocker:references/CONTRIBUTING.md:17:pipe-to-shell-pattern",
        "category": "blocker",
        "severity": "critical",
        "pattern": "Pipe to shell pattern",
        "file": "references/CONTRIBUTING.md",
        "line_start": 17,
        "line_end": 17,
        "snippet": "curl -LsSf https://astral.sh/uv/install.sh | sh"
      },
      {
        "id": "blocker:references/main_readme.md:12:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "references/main_readme.md",
        "line_start": 12,
        "line_end": 12,
        "snippet": "If you're new to working with the Claude API, we recommend starting with our [Claude API Fundamental"
      },
      {
        "id": "blocker:references/main_readme.md:26:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "references/main_readme.md",
        "line_start": 26,
        "line_end": 26,
        "snippet": "To avoid duplication of efforts, please review the existing issues and pull requests before contribu"
      },
      {
        "id": "blocker:references/README.md:12:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "references/README.md",
        "line_start": 12,
        "line_end": 12,
        "snippet": "If you're new to working with the Claude API, we recommend starting with our [Claude API Fundamental"
      },
      {
        "id": "blocker:references/README.md:26:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "references/README.md",
        "line_start": 26,
        "line_end": 26,
        "snippet": "To avoid duplication of efforts, please review the existing issues and pull requests before contribu"
      },
      {
        "id": "filesystem:scripts/memory_tool.py:307:python-shutil-operations",
        "category": "filesystem",
        "severity": "medium",
        "pattern": "Python shutil operations",
        "file": "scripts/memory_tool.py",
        "line_start": 307,
        "line_end": 307,
        "snippet": "shutil.rmtree(full_path)"
      },
      {
        "id": "filesystem:scripts/memory_tool.py:359:python-shutil-operations",
        "category": "filesystem",
        "severity": "medium",
        "pattern": "Python shutil operations",
        "file": "scripts/memory_tool.py",
        "line_start": 359,
        "line_end": 359,
        "snippet": "shutil.rmtree(self.memory_root)"
      },
      {
        "id": "blocker:scripts/memory_tool.py:49:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "scripts/memory_tool.py",
        "line_start": 49,
        "line_end": 49,
        "snippet": "ValueError: If path is invalid or attempts to escape memory directory"
      },
      {
        "id": "blocker:scripts/memory_tool.py:113:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "scripts/memory_tool.py",
        "line_start": 113,
        "line_end": 113,
        "snippet": "\"Valid commands are: view, create, str_replace, insert, delete, rename\""
      },
      {
        "id": "blocker:scripts/memory_tool.py:168:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "scripts/memory_tool.py",
        "line_start": 168,
        "line_end": 168,
        "snippet": "return {\"error\": f\"Cannot read {path}: File is not valid UTF-8 text\"}"
      },
      {
        "id": "blocker:scripts/memory_tool.py:261:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "scripts/memory_tool.py",
        "line_start": 261,
        "line_end": 261,
        "snippet": "\"error\": f\"Invalid insert_line {insert_line}. \""
      },
      {
        "id": "blocker:scripts/memory_tool.py:295:system-reconnaissance",
        "category": "blocker",
        "severity": "low",
        "pattern": "System reconnaissance",
        "file": "scripts/memory_tool.py",
        "line_start": 295,
        "line_end": 295,
        "snippet": "\"error\": f\"Invalid operation: Path '{path}' is not within /memories directory. \""
      },
      {
        "id": "metadata:SKILL.md:1:unscanned-file-invalid-frontmatter-manual-review",
        "category": "metadata",
        "severity": "low",
        "pattern": "Unscanned file (invalid_frontmatter) — manual review required",
        "file": "SKILL.md",
        "line_start": 1,
        "line_end": 1,
        "snippet": "[unscanned: invalid_frontmatter]"
      }
    ],
    "finding_verdicts": [
      {
        "id": "network:references/CONTRIBUTING.md:17:hardcoded-url",
        "verdict": "false_positive",
        "confidence": 0.95,
        "reason": "The URL is the official Astral install script for the uv package manager, shown in a documentation setup step. It is a well-known, legitimate host, not a data-exfiltration endpoint."
      },
      {
        "id": "network:references/CONTRIBUTING.md:27:hardcoded-url",
        "verdict": "false_positive",
        "confidence": 0.98,
        "reason": "This is a git clone of the official anthropics/anthropic-cookbook repository in a contributor setup guide. It is standard onboarding documentation with no security risk."
      },
      {
        "id": "env_access:references/CONTRIBUTING.md:117:python-environment-access",
        "verdict": "false_positive",
        "confidence": 0.95,
        "reason": "os.environ.get(\"ANTHROPIC_API_KEY\") is the recommended, secure way to read an API key from the environment. It is a documentation example, not credential harvesting; the key is not transmitted anywhere untrusted."
      },
      {
        "id": "env_access:references/CONTRIBUTING.md:117:generic-api-secret-keys",
        "verdict": "false_positive",
        "confidence": 0.95,
        "reason": "The line reads the user's own Claude API key from an environment variable in an example snippet. This is the documented best practice for handling secrets, not exposure or exfiltration of a key."
      },
      {
        "id": "sensitive:references/CONTRIBUTING.md:48:environment-file-access",
        "verdict": "false_positive",
        "confidence": 0.95,
        "reason": "cp .env.example .env is the standard scaffolding step for creating a local config file from a template. It reads no secrets and only copies an example file provided by the repository."
      },
      {
        "id": "sensitive:references/CONTRIBUTING.md:49:environment-file-access",
        "verdict": "false_positive",
        "confidence": 0.97,
        "reason": "This is a comment instructing the user to add their own API key to the .env file. It is documentation prose, not code that reads or transmits an environment file."
      },
      {
        "id": "blocker:references/CONTRIBUTING.md:17:pipe-to-shell-pattern",
        "verdict": "false_positive",
        "confidence": 0.9,
        "reason": "The pipe-to-shell command installs the uv package manager from Astral's official domain and is copied verbatim from upstream tooling documentation. While pipe-to-shell is generally discouraged, here it is benign vendor-provided setup instruction, not attacker-controlled code."
      },
      {
        "id": "blocker:references/main_readme.md:12:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.98,
        "reason": "The snippet is ordinary documentation prose pointing new users to a Claude API fundamentals guide. It performs no system reconnaissance; the pattern matched on benign English text."
      },
      {
        "id": "blocker:references/main_readme.md:26:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.98,
        "reason": "This is contributor guidance asking readers to review existing issues and pull requests. It is prose in a README, not code that enumerates the host system."
      },
      {
        "id": "blocker:references/README.md:12:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.98,
        "reason": "Identical documentation text pointing users to a Claude API fundamentals guide. No reconnaissance behavior; the pattern misfired on descriptive prose."
      },
      {
        "id": "blocker:references/README.md:26:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.98,
        "reason": "Contributor guidance prose about reviewing existing issues and pull requests. It is documentation, not code that gathers system information."
      },
      {
        "id": "filesystem:scripts/memory_tool.py:307:python-shutil-operations",
        "verdict": "false_positive",
        "confidence": 0.9,
        "reason": "shutil.rmtree(full_path) deletes a directory only after full_path has passed _validate_path, which confines every path to the /memories sandbox and rejects directory traversal. This is a scoped, validated delete inside a reference memory-tool implementation, not an arbitrary filesystem wipe."
      },
      {
        "id": "filesystem:scripts/memory_tool.py:359:python-shutil-operations",
        "verdict": "false_positive",
        "confidence": 0.88,
        "reason": "shutil.rmtree(self.memory_root) clears only the tool's own /memories storage directory in a clearly-documented, test-oriented helper. It cannot reach outside memory_root and carries an explicit warning; not a destructive attack on the host."
      },
      {
        "id": "blocker:scripts/memory_tool.py:49:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.97,
        "reason": "The snippet is a docstring line describing the ValueError raised on invalid paths. It is a defensive security control's documentation, not reconnaissance."
      },
      {
        "id": "blocker:scripts/memory_tool.py:113:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.97,
        "reason": "This is an error-message string listing the tool's valid commands. It is user-facing help text, not system enumeration."
      },
      {
        "id": "blocker:scripts/memory_tool.py:168:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.97,
        "reason": "An error message returned when a file is not valid UTF-8. It is benign error handling, not reconnaissance of the host."
      },
      {
        "id": "blocker:scripts/memory_tool.py:261:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.97,
        "reason": "An input-validation error message for an out-of-range insert_line value. It is defensive validation text, not system reconnaissance."
      },
      {
        "id": "blocker:scripts/memory_tool.py:295:system-reconnaissance",
        "verdict": "false_positive",
        "confidence": 0.97,
        "reason": "An error message enforcing that operations stay within /memories. This is part of the path-traversal defense, the opposite of malicious behavior."
      },
      {
        "id": "metadata:SKILL.md:1:unscanned-file-invalid-frontmatter-manual-review",
        "verdict": "false_positive",
        "confidence": 0.9,
        "reason": "Manual review of SKILL.md confirms it is a benign educational skill definition. It failed static scanning only because a 'TRANSLATED CONTENT:' line precedes the YAML frontmatter delimiter; the content itself is safe."
      }
    ],
    "semantic_findings": [],
    "critical_findings": [],
    "high_findings": [],
    "medium_findings": [],
    "low_findings": [
      {
        "title": "Unscanned file (invalid_frontmatter) — manual review required",
        "description": "[unscanned: invalid_frontmatter]",
        "locations": [
          {
            "file": "SKILL.md",
            "line_start": 1,
            "line_end": 1
          }
        ],
        "review_kind": "security",
        "source_category": "metadata",
        "source_severity": "low",
        "confidence": 1,
        "confidence_reasoning": "Force-confirmed metadata/low static finding; AI false-positive verdict rejected."
      }
    ],
    "dangerous_patterns": [],
    "files_scanned": 10,
    "total_lines": 970,
    "audit_model": "claude",
    "audited_at": "2026-07-18T09:19:17.821Z",
    "agent_auto_install_policy": "confirmation_required",
    "manual_install_policy": "allowed"
  },
  "content": {
    "user_title": "Learn Claude API With Cookbook Examples",
    "value_statement": "Developers often struggle to find working, up-to-date code patterns for the Claude API. This skill provides curated examples, tutorials, and best practices so you can build Claude-powered applications faster.",
    "seo_keywords": [
      "Claude",
      "Codex",
      "Claude Code",
      "Claude API",
      "Anthropic cookbook",
      "tool use",
      "RAG",
      "prompt engineering",
      "AI agents",
      "multimodal"
    ],
    "actual_capabilities": [
      "Provides basic Claude API message request examples in Python.",
      "Demonstrates tool use and function calling patterns.",
      "Shows multimodal usage such as image and vision analysis.",
      "Includes a secure reference implementation of the Claude memory tool.",
      "Documents contributor setup and development workflows.",
      "Covers retrieval augmented generation and third-party integration patterns."
    ],
    "limitations": [
      "It is a reference and documentation skill, not an executable application.",
      "Some code examples reference older model identifiers that may need updating.",
      "It does not include the full Anthropic cookbook, only selected translated examples.",
      "It does not manage API keys or credentials for the user."
    ],
    "use_cases": [
      {
        "title": "Learn Claude API basics",
        "description": "A developer new to the Claude API reads the examples to write a first message request.",
        "target_user": "Beginner developers"
      },
      {
        "title": "Implement tool use",
        "description": "An application engineer copies the function-calling patterns to add tools to an existing agent.",
        "target_user": "Application engineers"
      },
      {
        "title": "Build a memory backend",
        "description": "An AI engineer adapts the secure memory tool handler to persist context across sessions.",
        "target_user": "AI engineers"
      }
    ],
    "prompt_templates": [
      {
        "title": "First API call",
        "prompt": "Show me a minimal Python example that sends a single message to the Claude API and prints the response.",
        "scenario": "Getting started with the Claude API"
      },
      {
        "title": "Add a tool",
        "prompt": "Using the cookbook examples, help me define a get_weather tool and handle the tool_use response from Claude.",
        "scenario": "Adding function calling to an application"
      },
      {
        "title": "Analyze an image",
        "prompt": "Show me how to send an image to Claude and ask it to describe what the image contains.",
        "scenario": "Building a multimodal feature"
      },
      {
        "title": "Persist memory securely",
        "prompt": "Explain how the memory tool handler validates paths, then help me extend it to store user preferences under a per-user directory.",
        "scenario": "Building a stateful agent with persistent memory"
      }
    ],
    "output_examples": [
      {
        "input": "Show me a minimal Python example that sends a message to the Claude API.",
        "output": "The skill returns a short walkthrough that creates an Anthropic client, calls the messages endpoint with a model and a user message, and reads the text from the response content."
      },
      {
        "input": "How does the memory tool prevent directory traversal?",
        "output": "The skill explains that every path is passed through a validation step that requires a /memories prefix and confirms the resolved path stays inside the memory root, rejecting any attempt to escape the sandbox."
      }
    ],
    "best_practices": [
      "Store your Claude API key in an environment variable rather than hardcoding it.",
      "Update model identifiers in examples to the latest supported Claude model before use.",
      "Review any install script before piping it directly to a shell."
    ],
    "anti_patterns": [
      "Do not paste your API key directly into example code that may be shared.",
      "Do not run the memory tool handler outside its intended /memories sandbox directory.",
      "Do not treat the examples as production-ready without adding error handling and validation."
    ],
    "faq": [
      {
        "question": "What is this skill for?",
        "answer": "It provides code examples and tutorials for building applications with the Claude API."
      },
      {
        "question": "Is this an official Anthropic skill?",
        "answer": "It is a community translation of examples from the official Anthropic cookbooks. It is not maintained by Anthropic."
      },
      {
        "question": "Which programming language do the examples use?",
        "answer": "The examples are written in Python."
      },
      {
        "question": "Does the skill store my API key?",
        "answer": "No. The examples read the API key from an environment variable that you set yourself."
      },
      {
        "question": "Is the memory tool safe to use?",
        "answer": "The memory tool handler validates all paths and confines operations to a sandbox directory, but you should still review it before production use."
      },
      {
        "question": "Are the code examples always up to date?",
        "answer": "Some examples may reference older model names. Update the model identifier to a current Claude model before running them."
      }
    ]
  },
  "file_structure": [
    {
      "name": "references",
      "type": "dir",
      "path": "references",
      "children": [
        {
          "name": "capabilities.md",
          "type": "file",
          "path": "references/capabilities.md",
          "lines": 20
        },
        {
          "name": "CONTRIBUTING.md",
          "type": "file",
          "path": "references/CONTRIBUTING.md",
          "lines": 226
        },
        {
          "name": "index.md",
          "type": "file",
          "path": "references/index.md",
          "lines": 33
        },
        {
          "name": "main_readme.md",
          "type": "file",
          "path": "references/main_readme.md",
          "lines": 70
        },
        {
          "name": "multimodal.md",
          "type": "file",
          "path": "references/multimodal.md",
          "lines": 68
        },
        {
          "name": "patterns.md",
          "type": "file",
          "path": "references/patterns.md",
          "lines": 21
        },
        {
          "name": "README.md",
          "type": "file",
          "path": "references/README.md",
          "lines": 70
        },
        {
          "name": "third_party.md",
          "type": "file",
          "path": "references/third_party.md",
          "lines": 41
        },
        {
          "name": "tool_use.md",
          "type": "file",
          "path": "references/tool_use.md",
          "lines": 57
        }
      ]
    },
    {
      "name": "scripts",
      "type": "dir",
      "path": "scripts",
      "children": [
        {
          "name": "memory_tool.py",
          "type": "file",
          "path": "scripts/memory_tool.py",
          "lines": 364
        }
      ]
    },
    {
      "name": "SKILL.md",
      "type": "file",
      "path": "SKILL.md",
      "lines": 315
    }
  ]
}
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 323kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 323kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 323kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 323kUpdated Claude CodeCodexVerifiedInstall