Automattic/wordpress-atmosphere / .agents/skills/review
review
UnverifiedShared base for reviewing a diff, a branch or a PR in this plugin. How to read the diff, which repo guidelines to load for the changed paths, and the shape of a finding. Used by the code-review and security-audit agents.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 7.1 KB |
| guidelines/accessibility.md | markdown | 781 B |
| guidelines/authorization.md | markdown | 2.8 KB |
| guidelines/content-exposure.md | markdown | 6.8 KB |
| guidelines/cron.md | markdown | 1.9 KB |
| guidelines/hooks.md | markdown | 2.3 KB |
| guidelines/oauth.md | markdown | 3.6 KB |
| guidelines/output.md | markdown | 610 B |
| guidelines/records.md | markdown | 2.1 KB |
| guidelines/remote-fetch.md | markdown | 5.3 KB |
| guidelines/supply-chain.md | markdown | 582 B |
| guidelines/uninstall.md | markdown | 2.0 KB |
| guidelines/well-known.md | markdown | 1.4 KB |
| judge.md | markdown | 1.2 KB |
guidelines/records.md
View on GitHub# Records
What goes into `app.bsky.feed.post` and `site.standard.document`, and what building it costs. Who may see a record is in `content-exposure.md`; whether it matches the Lexicon is the spec-check agent's job.
Applies to: `includes/class-publisher.php`, `includes/transformer/**`, `includes/content-parser/**`, `includes/class-mention.php`, `includes/class-link.php`
## Past issues
- **Per-call egress cap read as per-publish** (#165 review, residual): `MAX_RESOLVED_HANDLES` bounds one `Facet::extract()` call, but a teaser thread re-enters `extract()` per entry and multiplies the bound (about 5 × 20 lookups per publish, each up to DNS plus a 10 s HTTPS fetch).
## Building records
- Post content is sanitized before transformation.
- The `applyWrites` batch cannot be steered into writing unexpected records or collections.
- TID generation is not predictable and does not collide.
- Facet detection (`class-facet.php`) cannot inject malicious links or mentions.
- The bsky post transformer respects Bluesky's character and image limits.
## Mention resolution budget
Check four things together:
1. The worst case per publish or request, not per function call: add up every `Facet::extract()` call the publish makes (thread entries and retries multiply a per-call cap).
2. Misses are cached as well as hits, or a dead handle is resolved again on every call.
3. Commenter-controlled content never triggers resolution unless the site opts in with a boolean filter that is cast with `(bool)`.
4. `Resolver::is_valid_handle()` rejects reserved and numeric TLDs before any network I/O (see "Handle syntax" in `remote-fetch.md`).
The fix for attacker-chosen DNS is at the threat-model layer (gating the commenter path, an allow list, DoH with a timeout), not in the syntax regex.
## Publish and display parity
Whatever decides "this `@handle` is a mention" or "this is a link" is shared between the front-end linkifier and the facet extractor. Two tokenizers drift: notifications without links, links without notifications, including `#mention` facets (real Bluesky notifications) for handles inside `<code>`, `<pre>` or `<a>` that the site's own display refuses to link.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |