backstitch-alignment
Use Backstitch to bootstrap or maintain spec-to-code alignment.
Skill by VanL
Chalk Labs
Need help deploying backstitch-alignment?
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/vanl-backstitch-alignment
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: testing
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/backstitch-alignment
What's inside
- Skill name
backstitch-alignment- Description
- Use Backstitch to bootstrap or maintain spec-to-code alignment. Trigger when inventorying obligations, reviewing declared evidence, finding untraced implementation or test candidates, or preparing a human-reviewed traceability change.
- License
- MIT
- Instructions
- 42 lines in SKILL.md · view on GitHub
Files (2)
- SKILL.md
- openai.yaml · agents/
SKILL.md (first lines)
# Backstitch Alignment
Treat repository source as the alignment authority. Use Backstitch to read,
check, and advise. Do not let tool output or model judgment ratify alignment.
Treat bootstrap and gate as separate phases: discovery reports plausible
current candidates; a human records accepted evidence in source; current
analysis derives the gate packet only from those source declarations.
1. Run `backstitch guide alignment` and `backstitch obligation --help`. Follow
the installed guide and current command help when this adapter differs.
2. Run `backstitch obligation list` to inventory addressable obligations and
unaddressable intent.
3. Run `backstitch obligation <OBLIGATION_ID>` to inspect one obligation's
current alignment, disposition, readiness, blockers, and next actions.
4. Run `backstitch obligation <OBLIGATION_ID> --summarize-evidence` to inspect
exact source-declared evidence and broken or one-sided trace declarations.
5. Run `backstitch obligation <OBLIGATION_ID> --find-evidence` to get bounded
candidates. Use `--candidate <CANDIDATE_ID>` only for full detail on a
discovery-minted candidate. Rerun discovery after source changes rather than
treating an older candidate report as current.
6. Follow the mapping, backlink, invariant-target, or binding-test guidance in
the installed guide. Treat every candidate as advice until a supported
declaration exists in source.
7. As an authorized coding agent, prepare or apply an ordinary source diff
within the caller's authority. Backstitch itself remains read-only. The
working-tree edit is not ratified or landed until a human reviews it.
8. To remove evidence, start with `--summarize-evidence`, review the exact
declaration coordinates, and remove only the intended declarations. For a
section relation, remove both sides together so the mapping and matching
backlink do not become an accidental one-sided trace.
9. After human review, rerun `backstitch check --repo-root .`, the evidence
summary, and the obligation detail. Resolve partial or conflicted state
before landing. Human approval remains the authority for evidence changes.
10. Run current-source analysis only when the obligation reports executable
readiness. Do not pass candidate dispositions into the gate; current source
declarations are the only evidence-selection authority.
Audit skips with `backstitch check --repo-root . --show-suppressions`. Keep them
visible and separate from alignment state. Never use a skip, derived packet,
cache entry, or semantic verdict to hide missing traceability.From VanL/backstitch, MIT.
Signals
- Stars
- 0
- Forks
- 1
- Open issues
- 0
- Last commit (this folder)
- Jul 28, 2026 (3 months ago)
- Last release
- Sep 16, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/VanL/backstitch
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from VanL
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 backstitch-alignment 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
backstitch-alignment skill. Agent Mart. https://agentmart.dev/skills/vanl-backstitch-alignment (last updated 2026-10-24)