ase-repo-review
Perform a holistic, human-reviewer-style critique of the currently staged Git changes and emit an approve/reject verdict with prioritized, severity-tagged, line-cited findings. Use when the user wants
SKILL.mdplugin/skills/ase-repo-review/SKILL.md
GitHub@${CLAUDE_SKILL_DIR}/../../meta/ase-control.md @${CLAUDE_SKILL_DIR}/../../meta/ase-skill.md @${CLAUDE_SKILL_DIR}/../../meta/ase-getopt.md
<purpose name="ase-repo-review"> Review Changes </purpose><expand name="getopt" arg1="ase-repo-review" arg2="--severity|-S=(AUTO|LOW|MEDIUM|HIGH) --thorough|-t --staged --unstaged --branch|-b= --commit|-c="> $ARGUMENTS </expand>
<objective> Review the currently staged (and/or unstaged) Git changes, the changes of a branch against its base, or a single commit the way an *experienced human reviewer* would - judging them *holistically* against the change's *own intent* and against *correctness*, *design fit*, *clarity*, *robustness*, *propagation*, freedom of *residue*, and *project-convention conformance* - and emit a single *approve / request-changes verdict* backed by *prioritized*, *severity-tagged*, *line-cited* findings. This is a *synthesizing critique*, not a mechanical scan: it complements `ase-code-lint` (mechanical quality), `ase-code-analyze` (logic/semantics), and `ase-repo-diff` (intent narrative and risk). </objective>Procedure
<flow>- <step id="STEP 1: Determine Change Set">
- <if condition="<getopt-option-commit/> is not empty and (<getopt-option-branch/> is not empty or <getopt-option-staged/> is equal `true` or <getopt-option-unstaged/> is equal `true`)">
Only output the following <template/> and then *STOP* immediately:
<template>
⧉ **ASE**: ✪ skill: **ase-repo-review**, ▶ ERROR: option `--commit` cannot be combined with `--branch`, `--staged`, or `--unstaged`
</template>
</if>
Set <changes/> to the uncommitted changes to review: to
staged and unstagedif both <getopt-option-staged/> and <getopt-option-unstaged/> are equaltrue, tostagedif only <getopt-option-staged/> is equaltrue, tounstagedif only <getopt-option-unstaged/> is equaltrue, and to empty otherwise. Here, theunstagedchanges always include the untracked files. - <if condition="<getopt-option-branch/> is not empty">
Determine the *base* of the branch:
Run the command
<template> ⧉ **ASE**: ✪ skill: **ase-repo-review**, ▶ ERROR: branch **<getopt-option-branch/>** does not exist </template>git rev-parse --verify --quiet "<getopt-option-branch/>^{commit}". If it fails, only output the following <template/> and then STOP immediately:Determine the candidate base branches by running the command
<template> ⧉ **ASE**: ✪ skill: **ase-repo-review**, ▶ ERROR: no base branch determinable for branch **<getopt-option-branch/>** </template>git for-each-ref --format="%(refname:short)" refs/heads/(taken exactly as given), excluding <getopt-option-branch/> itself. For each <candidate/>, run the commandgit rev-list --count "<candidate/>..<getopt-option-branch/>"to count the commits of the branch since its fork point from <candidate/>, and drop every candidate with a count of0, as it already contains the branch. Set <base/> to the remaining candidate with the smallest count (the nearest fork point); on a tie, prefer the checked-out branch (the output ofgit branch --show-current), and then the first candidate. If no candidate remains, only output the following <template/> and then STOP immediately:Run the command
git merge-base "<base/>" "<getopt-option-branch/>"and capture its output into <base-commit/>.Only output the following <template/>:
<template> ⧉ **ASE**: ✪ skill: **ase-repo-review**, ⎇ branch: **<getopt-option-branch/>**, ⎇ base: **<base/>** </template>
Determine whether there are changes at all by running the corresponding command(s) (taken exactly as given) and capturing their combined output - the bare list of changed file names - into <diff/>. This is a lightweight gate; the full diff is fetched by the sub-agent in STEP 2, so capturing only the file-name list here is sufficient:
<if condition="<getopt-option-branch/> is not empty and <changes/> is not empty"> Determine the *worktree* of the branch by running the command `git worktree list --porcelain` and setting <worktree-dir/> to the `worktree` path of the entry whose `branch` line is `refs/heads/<getopt-option-branch/>` (or to empty if there is none, as a branch not checked out anywhere carries no uncommitted changes).If <worktree-dir/> is not empty, set <scope>branch <getopt-option-branch/> since base commit <base-commit/> plus the <changes/> changes in worktree <worktree-dir/></scope> and run:
If <changes/> is equal
staged and unstaged:git -C "<worktree-dir/>" diff --name-only "<base-commit/>"git -C "<worktree-dir/>" ls-files --others --exclude-standardIf <changes/> is equal
staged:git -C "<worktree-dir/>" diff --cached --name-only "<base-commit/>"If <changes/> is equal
unstaged:git diff --name-only "<base-commit/>..<getopt-option-branch/>"git -C "<worktree-dir/>" diff --name-onlygit -C "<worktree-dir/>" ls-files --others --exclude-standard
From the output of the
ls-filescommand, skip every listed entry below the.ase/directory -- it carries ASE's own state, and hence is never part of the user's change set.Otherwise, set <scope>branch <getopt-option-branch/> since base commit <base-commit/></scope> and run:
git diff --name-only "<base-commit/>..<getopt-option-branch/>"</if> <elseif condition="<getopt-option-branch/> is not empty"> Set <scope>branch <getopt-option-branch/> since base commit <base-commit/></scope> and run:git diff --name-only "<base-commit/>..<getopt-option-branch/>"</elseif> <elseif condition="<getopt-option-commit/> is not empty"> Set <scope>commit <commit-id/></scope> and run:git show --format= --name-only --diff-merges=first-parent "<commit-id/>"</elseif> <elseif condition="<changes/> is equal `staged and unstaged`"> Set <scope>staged and unstaged</scope> and run:git diff --name-only HEADgit ls-files --others --exclude-standardFrom the output of the second command, skip every listed entry below the
.ase/directory -- it carries ASE's own state, and hence is never part of the user's change set. </elseif> <elseif condition="<changes/> is equal `unstaged`"> Set <scope>unstaged</scope> and run:git diff --name-onlygit ls-files --others --exclude-standardFrom the output of the second command, skip every listed entry below the
.ase/directory -- it carries ASE's own state, and hence is never part of the user's change set. </elseif> <else> Set <scope>staged</scope> and run:git diff --cached --name-only HEAD</else>- <if condition="<diff/> is empty"> Only output the following <template/> and then *STOP* immediately: <template> ⧉ **ASE**: ✪ skill: **ase-repo-review**, ▶ status: **no changes to review** </template> </if>
- <if condition="<getopt-option-commit/> is not empty and (<getopt-option-branch/> is not empty or <getopt-option-staged/> is equal `true` or <getopt-option-unstaged/> is equal `true`)">
Only output the following <template/> and then *STOP* immediately:
<template>
⧉ **ASE**: ✪ skill: **ase-repo-review**, ▶ ERROR: option `--commit` cannot be combined with `--branch`, `--staged`, or `--unstaged`
</template>
</if>
- <step id="STEP 2: Review Investigation">
<if condition="<ase-project-boxing/> is equal `black`">
The project source artifacts are classified as a black box, so the user does not want the changes scrutinized or their findings surfaced. Skip the entire review investigation: do not invoke the
</if>Agenttool and do not read any change, set <findings/> to the empty list, set <hidden/> to0, set <verdict/> toSKIPPED (boxing: black), set <summary/> to a one-line neutral restatement of the change intent derived solely from STEP 1, and proceed directly to STEP 3.First, use the following <template/> to give a hint on this step:
<template> <ase-tpl-bullet-secondary/> **REVIEW INVESTIGATION** </template>Dispatch the review investigation to a sub-agent via the
Agenttool so that no investigation details leak into the user-visible transcript. The sub-agent performs the silent reading, the read-only repository probing, and the critique; only its final structured return value is consumed here.Set <mode>thorough</mode> if <getopt-option-thorough/> is equal
true, and <mode>focused</mode> otherwise.For this, invoke exactly once the tool:
Agent( description: "Review Investigation", subagent_type: "ase:ase-repo-review", prompt: "Review in <mode/> mode the changes of scope: <scope/>", run_in_background: false )Parse the single result message of the
Agenttool as a JSON object, set <summary/> to itssummaryfield (a single crisp sentence reconstructing the change's intent), and set <findings/> to itsfindingsfield (a list).Then derive the overall <verdict/> from <findings/>: set <verdict/> to
REJECT - DEMANDS CHANGESif any finding in <findings/> has aseverityfield ofHIGH; otherwise set <verdict/> toAPPROVE. The verdict is derived before the severity floor below, so the floor only affects which findings are rendered, never the verdict.Then determine the effective severity floor <floor/> and its origin <floor-origin/>: define the ordinal rank
<if condition="<getopt-option-severity/> is equal `AUTO`"> Set <floor>MEDIUM</floor> and <floor-origin>boxing: grey</floor-origin> if <ase-project-boxing/> is equal `grey` (grey boxing surfaces only *material* findings of severity `MEDIUM` and above), and <floor>LOW</floor> and <floor-origin>default</floor-origin> otherwise. </if> <else> Set <floor><getopt-option-severity/></floor> and <floor-origin>--severity <getopt-option-severity/></floor-origin>: an *explicitly* given severity floor always *wins* over the boxing. </else>LOW=1,MEDIUM=2,HIGH=3.Then apply the effective severity floor <floor/>: Keep a finding in <findings/> if and only if its
severityfield isACCEPTEDorrank(severity)is greater than or equal torank(<floor/>); drop all other findings, but set <hidden/> to the number of dropped findings. With the floorLOW, all findings are kept.ACCEPTEDfindings are never dropped. The floor affects only which findings are rendered, never the <verdict/> derived above.You MUST NOT output anything else in this STEP 2.
</step> - <step id="STEP 3: Verdict and Findings">
Use the following <template/> to output the overall review in <verdict/> and the reconstructed intent <summary/>:
<template><ase-tpl-bullet-signal/> REVIEW VERDICT: <verdict/>
<ase-tpl-bullet-normal/> CHANGE INTENT: <summary/>
</template>You MUST NOT output anything else in this STEP 3.1.
- <if condition="<findings/> is empty">
Only output the following <template/> - with <note/> set to `the
change was not reviewed` if <verdict/> starts with `SKIPPED`, to
`no finding reaches severity <floor/>` if <hidden/> is greater
than `0`, and to `the change is clean` otherwise - and then
*SKIP* substep 3 of this STEP 3:
<template>
<ase-tpl-bullet-normal/> NO FINDINGS: <note/>, nothing to flag.
</template> </if> - <if condition="<findings/> is NOT empty">
Sort the findings by <severity/> from highest to lowest in the
fixed order `HIGH`, `MEDIUM`, `LOW`, `ACCEPTED`. Within the same
severity, keep the order returned by the sub-agent.
Then render a three-column table with one row per finding by using the following output <template/>. For each finding, repeat the third line, set <severity/> to its
severityfield, set <dimension/> to itsdimensionfield, set <location/> to itslocationfield, and set <finding/> to itsfindingfield.In the <location/> column, mark up the
file:linereference as code (with backticks) and prepend it with▢- keep the sub-agent's own:N/:N-Mline citation intact and do not append any further line-count decoration.Because the <finding/> text is free-form Markdown, before emitting any row you MUST escape every literal
<template> | Severity | Dimension | Finding | | --------------- | ------------ | ----------------------- | | **<severity/>** | <dimension/> | <location/>: <finding/> | </template>|pipe character inside <location/> and <finding/> as\|so it cannot break the table column structure.Keep the overall report concise and brief. Do not output any further explanation.
</if> - <if condition="<hidden/> is greater than `0`">
Output the following <template/>, so that suppressed findings
are never mistaken for a clean change:
<template>
<ase-tpl-bullet-normal/> HIDDEN FINDINGS: <hidden/> finding(s) below severity <floor/> suppressed (<floor-origin/>).
</template> </if> Finally, give the closing hints by expanding the following (which, depending on the configured <ase-guidance-level/>, may expand into nothing and hence emit no output at all):
<if condition="<verdict/> starts with `REJECT`"> <ase-tpl-hint level="normal"> Use `/ase-task-edit` to derive a solution approach and plan for the findings which demand changes. </ase-tpl-hint> </if> <if condition="<hidden/> is greater than `0`"> Set <rerun>/ase-repo-review -S LOW</rerun>, append ` --staged` to <rerun/> if <getopt-option-staged/> is equal `true`, append ` --unstaged` to <rerun/> if <getopt-option-unstaged/> is equal `true`, append ` -b <getopt-option-branch/>` to <rerun/> if <getopt-option-branch/> is not empty, append ` -c <getopt-option-commit/>` to <rerun/> if <getopt-option-commit/> is not empty, and append ` -t` to <rerun/> if <getopt-option-thorough/> is equal `true`, so that the re-run reviews the same scope in the same mode. <ase-tpl-hint level="normal"> Use `<rerun/>` to re-run the review with all findings shown (a fresh review, so its findings can differ). </ase-tpl-hint> </if>