software-factory-reviewer
Reviewer role boundaries for Software Factory.
Skill by jack-michaud
Chalk Labs
Need help deploying software-factory-reviewer?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/jack-michaud-software-factory-reviewer
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: devops
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/software-factory-reviewer
What's inside
- Skill name
software-factory-reviewer- Description
- Reviewer role boundaries for Software Factory.
- License
- MPL-2.0
- Instructions
- 31 lines in SKILL.md · view on GitHub
Agent Skills spec issues
- body: contains an absolute path from the author's machine
Files (1)
- SKILL.md
SKILL.md (first lines)
# Software Factory Reviewer Verify artifacts independently with read-only checks unless a separate task grants mutation authority. ## Remote Sprite Development Review For reviews of existing remote Sprite work, load `remote-sprite-development` and independently check the builder evidence package: target Sprite locality, pre/post checkpoints, remote changed files/diffs, verification commands/results, and rollback procedure. Reviewer scope is read-only unless a separate task explicitly grants mutation authority. Fail or block with `skill_context_failure` when the skill/contract is missing, `locality_violation` for local-only proof, and `evidence_gap` for insufficient command/diff/test evidence. ## Conditional disposable/test-profile validation review Use read-only/static review by default. Low-risk docs/comment-only changes, typo fixes, and changes where acceptance criteria can be fully verified from public source diffs should remain static-review-only unless PM or the human explicitly requires disposable validation. When PM requires disposable/test-profile validation, gate the source and handoff for all of the following: - PM explicitly named the validation as required and identified the behavior-risk trigger. - The trigger is appropriate: SOUL/profile behavior, skill behavior, Kanban protocol, profile install/delete/config/model/provider, role boundary, credential/env-var loading, remote-sprite workflow guidance, or prior production/meta failure. - Disposable validation is not being over-applied to low-risk work without a reason. - Builder/installer guidance uses randomly suffixed disposable profile names, canonical Hermes profile install/delete commands, non-secret evidence capture, and blocks rather than guessing on command shape or authority. - PM-required validation is not silently skipped; unavailable validation must produce a blocked task/comment with concrete evidence. - Cleanup/pruning is represented as a gated follow-up after rollout/docs evidence is preserved, unless explicit human retention is recorded. - The chain limits repeated remediation to at most two iterations before orchestrator/human escalation. Precedent checks: disposable install evidence should include root distribution metadata such as `distribution.yaml` verification when relevant and may use an approved fallback executable if the expected local Hermes wrapper is broken; evidence should name the fallback without leaking private state; approved non-secret disposable-validation artifacts should be preserved before cleanup; cleanup should use canonical profile deletion after rollout/docs evidence is preserved. ## Source-update review locality For source-update reviews, verify against the PM Source Map and builder handoff. Inspect the named standard worktree/path, such as `/home/sprite/worktrees/<repo-name>/<task-id>-<short-slug>`, or the public remote coordinates named in the handoff. Do not use private installed profile state, secrets, logs, memories, raw Kanban databases/workspaces, or runtime profile directories as source truth. Require evidence for repo URL, local repo/worktree path, branch, commit hash, changed files, diff or diff-stat, validation output, target-profile coverage, publish/push status, and whether superproject submodule pointers or generated profile repos changed. If commits are not pushed, require either an approved source-controlled patch/diff bundle or a PM-named reviewer-accessible standard local worktree path. Block with the exact missing public coordinate or unblock condition when source authority, path access, branch/worktree authority, or validation evidence is ambiguous.
From jack-michaud/software-factory-reviewer-profile, MPL-2.0.
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- May 11, 2026 (5 months ago)
- Last release
- May 9, 2026 (5 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MPL-2.0
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from jack-michaud
- 0.474score
remote-sprite-development
jack-michaud
Use when a Software Factory worker task must mutate or verify an existing remote Fly.io Sprite while preserving locality, checkpoint, evidence, rollback, and…
SkillCFreshness: CClaude Code +9 - 0.471score
software-factory-kanban-workflow
jack-michaud
Single-profile Software Factory Kanban workflow with durable task context, reminder propagation, bounded remediation retries, escalation, and shortcut…
SkillCFreshness: CClaude Code +9 - 0.470score
remote-sprite-development
jack-michaud
Use when Software Factory production PM, builder, or reviewer tasks need to mutate or verify an existing remote Fly.io Sprite via explicit remote commands…
SkillCFreshness: CClaude Code +9 - 0.469score
software-factory-publisher
jack-michaud
Publisher role boundaries for Software Factory generated public profile repositories.
SkillCFreshness: CClaude Code +9 - 0.469score
software-factory-evidence-report
jack-michaud
Generate Software Factory PM evidence HTML/JSON reports from public Kanban task surfaces.
SkillCFreshness: CClaude Code +9 - 0.466score
software-factory-pm
jack-michaud
PM role boundaries for Software Factory.
SkillCFreshness: CClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get software-factory-reviewer running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
software-factory-reviewer skill. Agent Mart. https://agentmart.dev/skills/jack-michaud-software-factory-reviewer (last updated 2026-10-24)