Skip to content

docs(#581): document review pipeline propagation requirement #583

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
fullsend-ai-coder wants to merge 1 commit into
base: main
Choose a base branch
from
Open

docs(#581): document review pipeline propagation requirement #583

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 39 additions & 0 deletions AGENTS.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -111,3 +111,42 @@ following fields are part of the skill specification:
These fields are defined by the skill spec. A field's first appearance
in a skill file in this repo is not a novel pattern and should not be
flagged as a code-organization concern.

## 8. Review pipeline propagation

The review system has two layers that must stay in sync:

- **`skills/code-review/SKILL.md`** defines review dimensions and
mandatory-finding rules for standalone use (e.g., when a developer
invokes `/code-review` directly).
- **`skills/pr-review/SKILL.md`** orchestrates parallel sub-agents for
PR reviews. Each sub-agent receives only its own definition file from
`skills/pr-review/sub-agents/` — it never sees
`code-review/SKILL.md`.

Because sub-agents are context-isolated, a rule added to
`code-review/SKILL.md` has no effect during orchestrated PR reviews
unless the corresponding sub-agent definition also contains a matching
procedural section.

### Propagation rule

@fullsend-ai-review fullsend-ai-review Bot Jul 31, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[low] logic-error

Step 1 of the propagation rule says the roster table 'maps categories to sub-agents.' The roster table's Dimensions column contains prose descriptions of review scope, not formal category identifiers. The actual category-to-dimension mapping is in the step 3a table in pr-review/SKILL.md.

When adding or modifying a review rule in `code-review/SKILL.md`:

1. Identify which sub-agent owns the rule's dimension. The sub-agent
roster table in `pr-review/SKILL.md` (section "Sub-agent roster")
maps categories to sub-agents.
2. Add a procedural section to the corresponding sub-agent file in
`skills/pr-review/sub-agents/`. The section must include the
category name, severity guidance, and file-pattern or content-
matching instructions so the sub-agent can independently detect

@fullsend-ai-review fullsend-ai-review Bot Jul 31, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[low] logic-error

Step 3 instructs adding a new category to 'the roster table,' but the roster table uses prose descriptions in its Dimensions column. New category identifiers should be added to the step 3a table (section 'Group prior findings by review dimension'), which the orchestrator's step 3a actually uses to route prior findings to sub-agents.

and evaluate the condition.
3. If the rule introduces a new category not listed in the roster
table's dimension mapping, add the category to the appropriate
row in the roster table in `pr-review/SKILL.md`.

Use the "Permission and role changes" section in
`skills/pr-review/sub-agents/security.md` as the reference model for
what a complete procedural section looks like: it specifies the
owned categories, the file patterns that trigger evaluation, the
evaluation criteria, and the severity thresholds.
Loading