一键导入
apply-findings
Fix the punch-list findings assigned to you, one module at one level, smallest sufficient edit, verify then land. Run by editor agents.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Fix the punch-list findings assigned to you, one module at one level, smallest sufficient edit, verify then land. Run by editor agents.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Review dimension for naming, idiom, comment debt, and AI tells against the codified per-language and prose standards. Invoked by reviewer agents over one module shard.
Recipe for writing C in the runtime, module placement and the ownership and control-flow discipline. Invoked by writer agents when a C unit is dispatched.
Recipe for writing Clojure, Functional Core / Imperative Shell at the function level (namespaces by domain), native wrappers, and the boundary discipline. Invoked when writing Clojure for the project.
Recipe for writing Elixir, pure functions in modules as the core and GenServer plus OTP as the shell, supervision trees, and NIF discipline for the C and Zig edge. Invoked when writing Elixir for the project.
Recipe for any English the agent produces in this project, commits, comments, docstrings, ADRs, design docs, skill bodies, changelog, error messages, project guides. Optimizes for terse, humanized prose with no em dashes and no AI tells. Invoke for any prose-writing activity.
Recipe for writing Zig, native bodies and edge wrappers, the allocator and lifetime discipline, and the hot-path rules. Invoked when writing or editing any Zig source.
| name | apply-findings |
| description | Fix the punch-list findings assigned to you, one module at one level, smallest sufficient edit, verify then land. Run by editor agents. |
| user-invocable | false |
Fix the punch-list items assigned to you (one module, one fix level), one commit per finding or tightly-related group. You are the sole mutator in a fix loop.
Per finding:
FAILED <id>: not reproducible, <why> rather than a
speculative edit.:correctness level do not also rename or refactor; that is the
factoring wave's job and it makes the diff unreviewable. At
:factoring or :style the finding defines the scope.references/architecture.md), return needs-cross-module <id>; do
not write the tempting local workaround. That is exactly the fix the
no-workarounds rule bans.write-tests) unless one already covers it; "the
reviewer found it" means the suite did not. For a persistence fix,
the regression exercises the store's transaction and history
behavior. For a native fix, the round-trip integration test that
compiles and calls real native code.Category: Imperative summary with jj (jj describe, then jj new) and fold it into the stack by change id
(jj squash or jj rebase). Never edit CHANGELOG.md or version
fields. RETURN your changelog lines; the round runner collects them.
Changelog lines follow write-changelog (user-visible effects only;
no line is normal for a style or factoring fix).If two of your findings conflict (one wants the code the other
removes), fix in punch-list order and note the second as overtaken:
FAILED <id>: overtaken by <first-id>.
Owns applying one module's findings at one level. The level order and
wave batching are run-review-round; scope that crosses modules is the
round runner's call. Reviewers are read-only; the verifier is
bash-heavy with no judgment; the editor is the sole source mutator.
LANDED <id> <n> commit(s) / needs-cross-module <id> / FAILED <id>: <first error>.