Skip to content

mxyhi/ok-skills / opencli/opencli-adapter-author

opencli-adapter-author

Verified

Build or improve an OpenCLI JS adapter using observed site behavior, existing adapters and browser API evidence. For ad-hoc browser interaction, see opencli-browser.

references/strategy-selection.md

View on GitHub
# Choosing an adapter strategy

Use the site's available contract and the task's needs to choose an implementation.
This reference is optional; no strategy note or classification artifact is required.

| Source | Useful when | Check |
| --- | --- | --- |
| Public API | A documented endpoint covers the task | Inputs, authentication, pagination and error semantics |
| Cookie-authenticated API | The endpoint works with the current account | Cookie/CSRF scope and real returned identity |
| Page state / SSR | The page already contains the required data | Field meaning, completeness and hydration timing |
| Semantic UI | The visible controls express the operation reliably | Stable selectors, target identity and observable result |
| Browser fetch | The request needs page origin or runtime context | Actual request parameters and response meaning |
| Network interception | The page produces a request that is difficult to reproduce | The captured response belongs to the intended action and input |

There is no universal API-over-DOM rule. A stable visible control can be easier to
maintain than an undocumented internal endpoint. Reuse an existing adapter or
shared helper when it already models the same operation.

## Gather facts as needed

```sh
opencli browser state --diagnostics --session research
opencli browser network --static --session research
opencli browser network --detail RequestKey --session research
```

Read `data` from the success envelope. Page diagnostics expose title, URL and
initial-state globals. Network evidence provides captured request keys, response
shapes and static source locations. `static` is a candidate; `network+static`
means both sources support the route. Neither proves business correctness.
The CLI does not assign a strategy score or prescribe the next action.

Check representative values against the visible page. For lists, distinguish a
true end of pagination from a failed or partial response. For writes, bind the
operation to the requested target and inspect its result; do not replay an action
whose previous outcome is unknown.

If an endpoint needs a runtime-generated signature, consider observing the page's
own request or using semantic UI. Use `references/api-discovery.md` for techniques
and `references/deep-recon.md` when the protocol requires further investigation.
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: 322kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 322kUpdated 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: 322kUpdated 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: 322kUpdated Claude CodeCodexVerifiedInstall