docs(rules): enforce research gates across workflows
This commit is contained in:
1 parent
d929c6df26
commit
21e3352456
1 file changed
+4
-3
@@ -4,10 +4,11 @@ Apply instructions in this order: direct user instructions; security and data-lo
|
|||||||
|
|
||||||
Preserve security, data, compatibility, and unrelated user work. Ask only when behavior, contracts, permissions, or irreversible choices are materially ambiguous. Otherwise choose the smallest reversible change.
|
Preserve security, data, compatibility, and unrelated user work. Ask only when behavior, contracts, permissions, or irreversible choices are materially ambiguous. Otherwise choose the smallest reversible change.
|
||||||
|
|
||||||
Use one applicable workflow per task. Treat skills and plugins as guidance subordinate to these rules. Continue work already authorized by the user; additional design approvals, commits, or delegation require a task-specific reason and the necessary authorization.
|
Use one applicable workflow per task, with mandatory research gates for every search within it. Treat skills and plugins as guidance subordinate to these rules. Continue work already authorized by the user; additional design approvals, commits, or delegation require a task-specific reason and the necessary authorization.
|
||||||
|
|
||||||
Use direct reads and `rg` for known paths, symbols, exact strings, error messages, and localized edits.
|
Before EVERY search or lookup, load and apply `research` and the matching specialist: `repository-research` for repository sources; `technical-research` for external technology sources. This includes direct reads, file discovery, exact matches, web, documentation, MCP, connector, tool discovery, and follow-up searches in every workflow. Known identifiers, familiar areas, small tasks, and earlier research are never exemptions. Reuse skill instructions already in context, but apply their checks to each search. If a required skill is unavailable, report it and still identify the question, inspect authoritative results, and state evidence gaps.
|
||||||
When `semantic-search` is installed, run it first for any question whose identifiers are unknown; fall back to `rg` with a brief note only if it fails.
|
Within that research workflow, use direct reads and `rg` for known paths, symbols, exact strings, error messages, and localized edits.
|
||||||
|
When `semantic-search` is installed, apply it before every repository search to select the route; run semantic retrieval first when identifiers are unknown, and use its documented fallback if retrieval fails or remains irrelevant after rephrasing.
|
||||||
|
|
||||||
Verify proportionally: run the cheapest check that detects likely failure; broaden for security, data loss, integration, or cross-component risk. Never claim unrun checks passed. Report failures and material gaps.
|
Verify proportionally: run the cheapest check that detects likely failure; broaden for security, data loss, integration, or cross-component risk. Never claim unrun checks passed. Report failures and material gaps.
|
||||||
|
|
||||||
|
|||||||
Reference in new issue
Block a user