Cobra Dev
Developmentcobra-devCobra Dev is an Agent Skill for development published in the carapace-sh/carapace repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Use when developing or debugging spf13/cobra internals — Command struct internals, ExecuteC dispatch, Find vs Traverse, flag set hierarchy, completion engine (__complete, getCompletions, ShellCompDirective), shell script generation (bash V1/V2, zsh, fish, powershell), flag group validation, global state, Levenshtein suggestions, or understanding the internal execution model. Triggers on: "cobra internals", "cobra ExecuteC", "cobra Find", "cobra Traverse", "cobra findNext", "cobra mergePersistentFlags", "cobra ParseFlags", "cobra getCompletions", "cobra __complete", "cobra completion engine", "cobra bash V2", "cobra flag resolution", "cobra flag groups internals", "cobra global state", "cobra EnablePrefixMatching", "cobra EnableCaseInsensitive", "cobra Levenshtein", "cobra stripFlags", "cobra legacyArgs", "cobra execute", "cobra preRun", "cobra postRun", "cobra mousetrap", "cobra templateFuncs".
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/carapace-sh/carapace/blob/HEAD/skills/cobra-dev/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/cobra-dev, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Cobra Development Reference
Reference for developing and debugging cobra internals. Covers the Command struct, execution flow, flag resolution, completion engine, and shell script generation.
Data Flow
os.Args[1:]
→ ExecuteC()
→ Root().ExecuteC() (always starts at root)
→ InitDefaultHelpCmd, initCompleteCmd, InitDefaultCompletionCmd
→ Find(args) or Traverse(args) (based on TraverseChildren)
→ findNext() (exact match → alias → prefix match)
→ execute(flags)
→ ParseFlags
→ ValidateArgs
→ PersistentPreRun chain (parent → child)
→ PreRun
→ ValidateRequiredFlags, ValidateFlagGroups
→ Run / RunE
→ PostRun
→ PersistentPostRun chain (child → parent)
Sub-Resources
Load the reference that matches your task. When in doubt, load multiple references.
| Keywords | Reference |
|---|---|
| Command struct, all fields, internal state, commands slice, parent, flags, pflags, lflags, iflags, parentsPflags, FParseErrWhitelist, CompletionOptions, commandCalledAs, ctx | references/command-struct.md |
| ExecuteC, execute, Find, Traverse, findNext, stripFlags, argsMinusFirstX, legacyArgs, command resolution, flag stripping, prefix matching, TraverseChildren | references/execute-flow.md |
| Flags, PersistentFlags, LocalFlags, InheritedFlags, LocalNonPersistentFlags, NonInheritedFlags, mergePersistentFlags, ParseFlags, flag set hierarchy, parentsPflags, globNormFunc, Flag lookup, persistentFlag lookup | references/flag-resolution.md |
| __complete, __completeNoDesc, getCompletions, checkIfFlagCompletion, ShellCompDirective, CompletionFunc, RegisterFlagCompletionFunc, flagCompletionFunctions, completion output format, directive handling, ValidArgsFunction, completion flow | references/completion-engine.md |
| bash V1, bash V2, zsh, fish, powershell, GenBashCompletion, GenBashCompletionV2, GenZshCompletion, GenFishCompletion, GenPowerShellCompletion, script architecture, V1 vs V2 comparison, completion script internals | references/shell-completions.md |
| flag groups, MarkFlagsRequiredTogether, MarkFlagsMutuallyExclusive, MarkFlagsOneRequired, ValidateFlagGroups, processFlagForGroupAnnotation, annotation storage, requiredAsGroupAnnotation, oneRequiredAnnotation, mutuallyExclusiveAnnotation, enforceFlagGroupsForCompletion | references/flag-groups-internals.md |
| global state, OnInitialize, OnFinalize, initializers, finalizers, EnablePrefixMatching, EnableCaseInsensitive, EnableCommandSorting, EnableTraverseRunHooks, templateFuncs, AddTemplateFunc, MousetrapHelpText, MousetrapDisplayDuration, CheckErr, Gt, Eq | references/global-state.md |
| suggestions, Levenshtein, findNext, prefix matching, case-insensitive, SuggestFor, SuggestionsMinimumDistance, DisableSuggestions, ld function, hasNameOrAliasPrefix, commandNameMatches | references/suggestions.md |
Quick Guide
- How does ExecuteC dispatch to the right command? → references/execute-flow.md
- What is the difference between Find and Traverse? → references/execute-flow.md
- How does the flag set hierarchy work? → references/flag-resolution.md
- How does the __complete hidden command work? → references/completion-engine.md
- How are shell completion scripts generated? → references/shell-completions.md
- How do flag group annotations work internally? → references/flag-groups-internals.md
- What global state does cobra maintain? → references/global-state.md
- How does the suggestion/typo system work? → references/suggestions.md
- What are all the Command struct fields? → references/command-struct.md
Cross-Project References
- For user-facing cobra topics (creating commands, defining flags, argument validation, lifecycle hooks, help/usage, completions API, Viper integration), use the cobra skill.
- For pflag internals (non-POSIX modes, carapace-pflag extensions), use the carapace-dev skill →
references/pflag.md.