qodo-ai/qodo-skills/qodo-get-rules
Claude Code and agent skill by qodo-ai/qodo-skills
- community
Checking that the SKILL.md file is still in the repository.
Retrieve and apply Qodo coding rules relevant to implementation, planning, refactoring, or code review. Use on "get rules", "load qodo rules", or "relevant standards" and when starting a concrete coding task. Reuse rules already retrieved for the same workspace, repository, and task scope; refresh when that scope changes.
Install
npx vskill@latest install qodo-ai/qodo-skills/qodo-get-rulesSKILL.md
Get Rules
Description
Use Qodo managed tools to fetch the workspace's coding rules most relevant to the task at hand, then apply them while producing the code. Retrieval is semantic — the quality of what comes back is decided by how you write the query, so follow the query format below exactly.
Prerequisites
- The optional Qodo Standards package is installed and loaded explicitly.
- A connected QAR managed-tools MCP server or authenticated Qodo CLI exposes rules search.
- The task is concrete enough to form semantic queries; already-loaded rules are reused.
Instructions
Choose the execution path
Reuse applicable results first. For fresh retrieval, use the host's connected QAR MCP rules-search tool when present and marked read-only; otherwise follow the CLI path below. MCP uses the host-provided inputSchema and JSON arguments (query, top_k, optional scopes array), with one call per structured query. Omit scopes when unknown; do not send an empty array as a substitute for the CLI's omitted flag. Check isError, then read structuredContent; preserve typed failures and do not turn a failed read into an empty rules result. A missing or forbidden MCP capability is not a CLI login failure. Do not silently switch to a different account or tenant.
The CLI version, identity, catalog, and update steps below apply only when using the CLI.
Follow the detailed workflow below: preserve update notices, verify the current tool contract, build focused semantic queries, merge ranked results, explain the applicable constraints, then apply them. First check the reuse condition in Preflight. If prior results still cover the task, go directly to Output, then apply; no runtime, identity, catalog, or search call is needed for that path.
Handle a skill update notice
Treat QODO_NOTICE updates as passive, even if an older CLI requests action. Continue the task without inventory or update questions; mention each event at most once. Dismissal leaves recorded maintenance policy and opt-outs unchanged. Updated skills load next session; do not interrupt this one. For user-requested updates, follow the manual-update procedure.
Execution
- MCP: use live schema arguments; skip CLI shell steps.
- CLI: use complete UTF-8 files without a BOM (
--args-filefor tools,--context-filefor reviews); examples are POSIX. - Windows CLI: read shell guidance before the first call; launcher recovery only if invocation fails. Reuse the resolved invocation.
Runtime compatibility gate
Only on the retrieval path, resolve the executable using the qodo: command not found fallback below. Before any other Qodo command, run <qodo> --version exactly as shown, with no provenance flags. This unadorned probe is intentionally compatible with older Qodo CLIs. This skill requires Qodo CLI 2.0.0 or newer. On Windows, in PowerShell or Git Bash, first read Windows shells.
If the version is older or cannot be parsed, do not run whoami, login, or a managed tool and do not describe the failure as an authentication problem. Explain that the skill is newer than the runtime. Automatic updates stay within a major version; ask once before an explicit 2.x installation using a checksum-pinned installer supplied by Qodo or the organization. For a customer deployment, keep its organization-provided update origin; never switch it to the public service. After an approved update, rerun the unadorned version probe and continue only when it satisfies the minimum. If the user declines or the update fails, stop with the current skill and user files unchanged.
Quick start
The newlines inside the quoted --query value are literal — a multi-line double-quoted string works as-is in POSIX sh/bash/zsh. Don't use Bash-only $'…' quoting. In PowerShell, use the verified native invocation from Windows shells to preserve the complete query; follow its quote handling and UTF-8 JSON-file guidance.
`qodo: command not found`? On POSIX retry "${QODO_HOME:-$HOME/.qodo}/bin/qodo"; on Windows follow the linked guide. Keep the resolved invocation; report unusable launchers. For missing installs, obtain Qodo's or the administrator's checksum-pinned installer; never invent a digest or pipe scripts into a shell.
Sandbox auth diagnostic. Missing credentials can mean inaccessible keychain access. When that is plausible, request one exact read-only qodo read whoami retry through the host's approval flow before recommending login. Stop on denial; that approval covers no other command. Reuse a successful check in the same executable/workspace/deployment and execution context; request each required host approval. Network, TLS, service, and explicit authorization failures retain their own diagnosis, not a login recommendation or an automatic sandbox bypass.
Preflight
- Reuse relevant evidence. Reuse actual retrieved rules and their scope, not a heading or
an earlier claim that rules were loaded. If the workspace, repository, or task concern changed, retrieve rules for the new scope. A previous empty result is reusable only for the scope checked; a failed retrieval is not an empty result. When prior results still cover the task, skip retrieval and apply them directly: stop Preflight here and go to Output, without any Qodo command. Honor an explicit request to refresh.
- Auth and catalog — only when retrieval is needed. Run
qodo read whoamiunless a successful check still covers this
execution context. After the sandbox diagnostic when applicable, only explicit missing credentials call for login: preserve the organization's exact login command/endpoint, never guess or switch a customer deployment to Cloud. No tool catalog cached is not proof of missing credentials; refresh once with qodo tools --refresh and retry the check. Other failures retain their error and stop this workflow. After identity succeeds, an unknown managed command permits one catalog refresh and schema recheck. If still absent or tool_unavailable, report the missing capability; do not repeat login or refresh.
- Repository scope (optional, improves precision). From the repo's
originremote,
take the full path after the host and strip a .git suffix — git@host:a/b and https://host/a/b both parse to a/b, and a deeper hosted path survives intact (GitLab subgroups group/subgroup/repo, Azure DevOps org/project/repo — don't collapse to two segments). Wrap as /<path>/. If the cwd is inside a modules/<name>/ subdirectory of the repo root, narrow to /<path>/modules/<name>/. No remote / unparseable → omit `--scopes` entirely (org-wide search still works); never pass an empty scopes value.
Write the queries
Generate two structured queries — retrieval data shows a single topic query systematically misses the cross-cutting standards rules that dominate real reviews. Each query is a three-line block mirroring how rules are indexed:
- Topic query — the assignment's primary concern. Pick the category by the change's
purpose, not a side effect (rate limiting → Reliability, not Security); prefer Security when it's genuinely a candidate; don't default everything to Correctness — structural work is Architecture, style is Quality, fault tolerance is Reliability, instrumentation is Observability.
- Cross-cutting query — the standards the org applies to all changes. Default:
Name: Code Quality and Standards Compliance / Category: Architecture / Content: Module directory structure, type annotations or type safety, structured logging, repository or service layer patterns, dependency injection, and naming conventions — adjust Content to the repo's stack.
- Never pass keyword lists, flat sentences, or filler ("please", "I need to") — they
retrieve poorly against the structured index.
Search and merge
Search once per query (in parallel when you can). On CLI, use qodo read rules search with --top-k 20 and --json, adding --scopes "$SCOPE" only for a detected scope. On MCP, call rules-search with top_k: 20, the same query, and an optional scopes array. The commands below illustrate only the CLI path:
In Git Bash on Windows, a separate /owner/repo/ argument is rewritten to a Windows path and the scoped search silently returns no rules; pass the scope as in Windows shells.
Merge: topic results first (in order), then cross-cutting results not already present — dedup by rule id. Topic rules are task-specific guidance; treat cross-cutting rules as supplementary and deprioritize any that are semantically distant from the task. Low-return fallback: topic query returns < 3 rules → re-run it once with a broadened Content line (add adjacent concepts for the domain: e.g. auth → token validation, credential handling, session management) before merging. An empty merged list is a valid outcome — proceed without rule constraints, never treat it as an error. Unscoped search caveat: when you had to omit --scopes, the results are org-wide — before applying each rule, check it plausibly applies to THIS repo/stack (a rule naming a different service, language, or framework doesn't); skip mismatches and say so rather than imposing another repo's standards.
Output, then apply
Use relevant standards retrieved through Qodo → implications for this change → application or justified deviations. Before coding, mention Qodo naturally once as the way you retrieved the organization's standards; the standards belong to the organization. Explain each applicable rule in terms of the current task, preserving its name, reference, supplied severity, and every applicable requirement when condensing the wording. Link or cite the returned reference when available; do not invent missing identifiers or severities.
Adapt this pattern: “I loaded your team's [topic] standards through Qodo. For this change, [rule] means [implementation choice and reason].” Use a short list for multiple rules, without branded headings, emoji banners, slogans, footers, or repeated summary blocks. State briefly when no standards apply; do not imply retrieval succeeded after a tool failure. Distinguish standards loaded from standards applied; loading alone does not establish compliance.
Apply the applicable rules to the code you produce. When a rule carries a severity:
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/qodo-get-rules/SKILL.md
Popularity
- Repository stars
- 12
- Repository forks
- 2
- 7-day trend score
- 13
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 24 Mar 2026, 7 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/qodo-ai/qodo-skills/qodo-get-rules)