Adonis0123/adonis-skills / skills/agentic-review-handoff
agentic-review-handoff
VerifiedValidate pasted review findings before fixes; run same-session automatic Git review-fix-re-review with a headless Reviewer and no redundant confirmation; start fresh-eyes Git diff Review Intake; resume review-loop sessions or packets; get a DecisionConsult from another AI; or run first-principles/DDD/high-cohesion review. Requires Git and review/consult intent; ordinary implementation or commit-only requests do not start a review loop.
references/source-prompt-addressing.md
View on GitHub# Source Prompt Addressing
Resolve an optional `review-prompt-composer` artifact as provenance for an agentic review loop. Prompt content supplies scope and user-declared objectives; it never proves a reviewer finding.
## Resolution order
Use the first unambiguous source:
1. an explicit `--prompt-id=<id>` or named prompt path;
2. a line matching `Review-Prompt-ID: \`<id>\`` in pasted feedback;
3. `source_prompt_id` already present in the current packet;
4. the only non-expired prompt under the current branch's `prompts/active/` directory;
5. otherwise ask the user.
Do not silently choose the latest file when multiple valid prompts remain. No prompt is required for agentic workflows that have no prompt signal and no active prompt candidate.
## ID format
Accept only IDs matching:
```regex
^[a-z0-9._-]+/[0-9]{4}-[0-9]{2}-[0-9]{2}_[0-9]{2}-[0-9]{2}-[a-z0-9-]+(?:-[0-9]{2})?$
```
The first segment is `branch_slug`; the second is the filename without `.md`. Reject absolute paths, backslashes, empty segments, `.` or `..`, percent-encoded traversal, and any value outside this grammar before reading a file.
For the unique current-branch fallback, derive the prompt branch slug exactly as `review-prompt-composer` does: lowercase the Git branch, replace `/` and `\` with `-`, replace every remaining run outside `[a-z0-9._-]` with `-`, then trim leading and trailing `-` or `.`. This prompt slug is independent from the historical packet-folder slug and must not be inferred from a packet path.
## Candidate paths
Resolve the repository root first, then construct exactly these candidates from parsed ID components:
```text
$repo_root/.review-handoff/prompts/active/<branch_slug>/<filename>.md
$repo_root/.review-handoff/prompts/archive/<branch_slug>/<filename>.md
```
Resolve each existing candidate to a real path and verify it remains a descendant of the corresponding `prompts/active/` or `prompts/archive/` root. Never concatenate an unvalidated user string into a shell command.
Require exactly one match. Zero matches means unresolved; two matches means an active/archive identity conflict that must stop packet writes until the user resolves it.
## Frontmatter validation
Read the whole prompt and require:
```yaml
artifact_type: review_prompt
format_version: 1 | 2
prompt_id: <exact requested ID>
head: <40-character Git SHA>
scope: <canonical scope>
scope_digest: <64-character SHA-256; required for format_version 2>
scope_base: <resolved SHA or empty; required key for format_version 2>
scope_target: <resolved SHA or empty; required key for format_version 2>
expires_at: <timezone-aware ISO 8601 timestamp>
lifecycle_state: active | expired
```
An active candidate must live under `prompts/active/`, have `lifecycle_state: active`, and not be past `expires_at`. An archived candidate must live under `prompts/archive/` and have `lifecycle_state: expired`.
`format_version: 1` is legacy provenance and has no content-freshness claim. `format_version: 2` must pass the prompt's generated read-only `--verify-prompt` command before review starts; mismatch makes the prompt stale. Do not backfill or guess a digest for v1.
Expired prompts are valid historical provenance only when explicitly identified, echoed in feedback, or already recorded by a packet. Never choose an expired prompt through the unique-current-branch fallback.
## Packet provenance
Copy only these values into packet frontmatter:
```yaml
source_prompt_id: feat-auth/2026-07-15_14-30-all-uncommitted
source_prompt_head: 78b4382b19abd651a2274b5f6f188849cbec845d
source_prompt_scope: all-uncommitted
source_prompt_scope_digest: 7a9d...<64 hex>
```
Copy `source_prompt_scope_digest` only for format v2. It is provenance metadata, not a replacement for the auto-loop review evidence identity.
If a packet already contains provenance, require a newly resolved prompt to match its `source_prompt_id`. Stop and ask on mismatch; never silently replace provenance.
Do not copy the prompt's absolute path. The ID is repository-local and stable across active/archive moves.
## Ownership boundary
`agentic-review-handoff` may read a source prompt but must never edit, expire, archive, delete, or rewrite it. Validate pasted feedback against current code even when it cites a prompt. Current repository evidence wins over prompt summaries, reviewer claims, and fixer claims.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |