git:20260829.d72b396 to git:20260910.361b07c

8 added, 1 removed. Audit A to A.

---

name: x-research

description: Route and run bounded public X research for posts, threads and replies, account audits, organic benchmarks, audience voice, creator discovery, campaign scouting, and market localization.

metadata:

postplus:

familyId: x

familyName: X

---

# X Research

Use this as the X research entrypoint when the user needs public evidence for a

marketing, creator, audience, competitor, campaign, or localization decision.

Apply shared rulebook and user-guidance rules from `postplus-shared`.

When a supported command completes but evidence is empty, sparse, noisy,

off-topic, or the wrong record type, apply the `postplus-shared` reference

`research-quality-recovery.md`; hard execution errors still fail fast.

## Reference Index

Always apply `references/shared-contract.md` before running a route.

| User asks for | Apply |

| --- | --- |

| Known account, competitor account, account comparison | `references/account-and-competitor-audit.md` |

| Public posts, hooks, formats, topics, competitor content | `references/organic-benchmark.md` |

| Replies, audience voice, objections, FAQ, phrase bank | `references/audience-voice.md` |

| Creators, experts, operators, partners, profile verification | `references/creator-discovery.md` |

| Launch, hashtag, mention, brand or competitor campaign | `references/campaign-scout.md` |

| Market, region, language, or cross-border comparison | `references/market-localization-scout.md` |

| Profile facts only | Apply `references/shared-contract.md`, then use `x-profiles` for known accounts or `x-user-search` for recall |

| Private accounts, DMs, hidden analytics, full follower graphs, ad spend, targeting, conversion, or ROAS | Stop or ask for a supported public X scope |

| Cross-platform request | Run only the X lane here; hand off other platforms |

## First Question

Ask one question only when the answer changes platform, route,

public/private boundary, first-pass scope, or output shape.

| Missing | Ask |

| --- | --- |

| Platform | `Which platform should I use first?` |

| X seed | `Please give one X seed: keyword, hashtag, handle, profile/post URL, competitor, campaign term, or creator criteria.` |

| Account audit without account | `Send the X handle/profile URL, or tell me the brand and closest competitors to use as recall seeds.` |

| Audience voice without thread | `Send a public X post URL, or give a query so I can shortlist one to three threads first.` |

| Creator discovery without seed | `Give one niche, topic, product, competitor, audience, or collaboration goal.` |

| Localization without market | `Which market or language should I compare first?` |

Do not ask for credentials, implementation choices, schemas, hidden

fields, private exports, cookies, or retry strategy.

## Run Discipline

1. Route the request with the table above.

2. Apply `references/shared-contract.md`.

3. Apply the selected workflow reference.

4. Run the narrowest collection chain that answers the first pass.

5. Stop after the first pass and report scope, evidence, limits, and next action.

Result record shapes are documented in the `postplus-shared` reference

`dataset-item-schemas.md`; consult it before processing results, and inspect one

real record only to verify.

Do not present a bounded sample as full-platform truth. Use only the public

filters shown by the selected route.

## Public Command Boundary

- Readiness diagnostics: `postplus doctor --skill x-research`.

- Inspect a route with `postplus research run <route> --help` only when its

semantic flags are not already clear.

- Run `postplus research run <route> --<query/handle/url flags> --limit <n>

--wait --output <result.json>`.

- Use only the semantic flags shown by the selected route.

- Resume a pending run with

`postplus research run --resume-from <result.json>`; never extract or

rewrite the opaque handle stored in that checkpoint.

- - Keep the first pass bounded. If a command fails, report the exact error and

+ - Keep the first pass bounded. If a command still fails after any bounded

+ recovery allowed by the executing PostPlus skill, report the exact error and

stop; do not silently swap sources or invent missing data.

- If the CLI returns a quote-confirmation challenge, run

`postplus quote confirm --json --challenge-file <challenge.json>` and retry

with the returned token.

```bash

postplus research run x-posts \

--wait \

--output ./result.json

```

+ **Bounded recovery:** Current PostPlus CLIs handle a compatible update and retry the command once when no agent-session restart is required. If an older CLI only reports that an update is required, run `postplus update` and retry once under the same condition. For a missing or invalid CLI session, run `postplus auth login` yourself; it opens the browser by default. Immediately share its exact URL as a clickable link for the user to **Connect**, then retry the original command once only after the CLI confirms success. Never ask the user to run the command or enter/compare a code, approve the connection for them, or automatically restart a cancelled/expired login. For a local usage rejection before remote work starts, use that command's `--help` to make one unambiguous correction from existing user input and retry once.

+ If PostPlus returns `postplus_cli_balance_required` with an `open_url` user action, give the user its exact label and URL and stop for account action. Do not invent a checkout link, claim whether provider work or charging occurred, or blindly resubmit after payment; continue from the command's documented status or checkpoint once the user confirms credits are available.

+ Otherwise stop and report the exact error. Never expose login polling secrets, resubmit an operation when remote work may have started, change user intent, bypass approval, switch providers, rewrite payloads, or make a second recovery attempt. After success, briefly say that PostPlus updated, using only the official update details PostPlus reported.