selvedge

selvedge.sh  ·  PyPI  ·  GitHub

Long-term memory for AI-coded codebases — including what was already tried and rejected.

Line attribution tells you who wrote something. Selvedge tells your agent what not to write next: the approaches this codebase already tried, reverted, and why. It's a git blame for AI agents, for the why rather than which model touched which line — captured live, by the agent, as the change happens, so nothing downstream has to guess at it.

Selvedge is a local MCP server. AI coding agents (Claude Code, Codex, Copilot, Cursor, Gemini CLI and Windsurf) call it as they work to log structured change events with reasoning. Your data stays in a SQLite file under .selvedge/ next to your code.

Local-first by default, team-server by choice, zero-LLM always.

---

Six months ago, your AI agent added a column called user_tier_v2. You don't know why. git blame points to a commit from claude-code with a generated message that says "Update schema." The session that made the change is long gone — and so is the prompt that produced it.

With Selvedge, you run this instead:

$ selvedge blame user_tier_v2

  user_tier_v2
  Changed     2025-10-14 09:31:02
  Agent       claude-code
  Commit      3e7a991
  Reasoning   User asked to add a grandfathering flag for legacy free-tier
              users during the pricing migration. Stores the original tier
              so we can backfill discounts without touching billing history.

That reasoning was captured by the agent in the moment — written into Selvedge from the same context that produced the change. Not inferred from the diff afterward by a second LLM. Not a hand-typed commit message.

---

---

Who Selvedge is for

Selvedge has two audiences. Same tool, same pip install, same SQLite file under .selvedge/. Different scale of pain.

Teams running long-term, AI-coded codebases. When the project is big enough that you (or someone else) will touch it again in six months, twelve months, three years — but most of it was written by an agent whose context evaporated the day each PR shipped. git blame tells you what changed. Selvedge tells you why — even after the agent session, the prompt template, the developer who asked for it, and the model version are all long gone. This is the original use case: production codebases, schema decisions, migrations, dependency changes that need an audit trail that survives turnover.

Solo developers using Claude Code on everyday projects. Side projects, weekend builds, the small internal tool you keep poking at. You don't need enterprise governance — you just need to remember why you (or your agent) did the thing you did yesterday, last week, last sprint. Run selvedge init once. Add four lines to your CLAUDE.md. From then on, selvedge blame is muscle memory — a way to talk to your past self when your past self was an LLM.

If you've ever come back to your own AI-built project and thought "what was this for again?", Selvedge is the missing piece.

---

The problem

Human-written code leaks intent everywhere — commit messages, PR descriptions, inline comments, the Slack thread that preceded it. AI-written code doesn't. The agent has perfect clarity about why it made each decision, but that context lives in the prompt and evaporates when the conversation ends.

Six months later, your team is debugging a schema decision with no trail. git blame tells you what changed and when. It can't tell you why.

Selvedge captures the why — live, by the agent itself, as the change is made. The diff is git's job. The why is Selvedge's.

---

What's new in v0.3.14

Explicit seven-day lookups work as documented.

The MCP prior_attempts tool now accepts window_minutes=10080, matching its seven-day default. Previously, sending that value explicitly failed validation because the time window incorrectly shared the 1,000-result pagination cap. The allowed window is 1–10,080 minutes; result limits remain capped at 1,000. No new dependencies, migrations or MCP tools.

---

What's new in v0.3.13

Keep the review context with a recorded rejection.

• Session-start summaries show when a decision has expired or needs manual review, including in the rejection section.
• Superseded decisions leave the revisit list without hiding unrelated decisions on the same path. History stays intact; reopening remains explicit.
• Feedback and correction guide: how reports become product decisions, and how to reopen a mistaken rejection using supersede.

---

Where Selvedge fits

AI agents call Selvedge as they work. Selvedge captures the why into a durable, queryable store and emits it back out — as Agent Trace records for cross-tool readers, as observability metadata that links into Sentry/Datadog stack traces, and as compliance artifacts for SOC 2 and EU AI Act audits.

Selvedge does not replace git (line-level what/when), PR review tools (review-time quality), agent observability (LLM call traces), or general-purpose code-host AI features. It sits between them — the provenance-as-first-class-citizen layer that everything else references.

---

How Selvedge compares

There's a fast-growing "git blame for AI agents" category. Here's where Selvedge fits — and where it deliberately doesn't.

	Rejected paths	Reasoning source	Granularity	Mechanism	Grouping	Storage
Selvedge	Queryable — prior_attempts returns tried → reverted → re-opened	Captured live, by the agent in the same context that produced the change	Entity — DB column, table, env var, dep, API route, function	MCP server — agent calls it as work happens	Changesets — named feature/task slugs across many entities	SQLite, zero deps
OpenLore	Purged — rejected is an inactive status, dropped from the queryable store after each decision sync (the annotation survives in the synced spec markdown)	Derived — tree-sitter static analysis of code state, plus commit-gated decision notes	AST node (18 languages + 12 IaC)	MCP server — one-time index + commit-time certificates	Call-graph edges	SQLite graph in .openlore/
AgentDiff (sunilmallya)	None	Inferred post-hoc by Claude Haiku from the diff at session end	Line	Claude Code lifecycle hooks → local daemon	Session/task	JSONL on disk
AgentDiff (codeprakhar25)	None	ed25519-signed cross-agent provenance	Line	Per-agent editor hooks + git hooks (sign at commit)	None	Signed traces in git refs
Origin	None — rework flags reverted AI code post-hoc, without rationale	Prompt receipts, captured live per turn	Line	Agent lifecycle hooks + git post-commit hook	None	Git notes + sessions branch
Git AI	None	Attribution metadata	Line	Agent-invoked checkpoint → Git notes at commit	None	Git notes
BlamePrompt	None	Prompt receipts — prompt, cost, tools; no stated rationale	Line	Agent-lifecycle hooks + post-commit hook	None	Git notes

Why "rejected paths" matter — the one that isn't copyable. The expensive failure isn't forgetting why a column exists. It's an agent confidently re-implementing something the team already killed for a good reason, six months after everyone who knew that left the context window. None of the line-attribution tools above surface rejected paths at all, and it isn't a feature gap they can close in a release — a line-oriented store has no notion of an entity that persisted across a try → revert → retry cycle. See docs/demos/prior-attempts.md.

Why determinism matters. Selvedge's reasoning is the agent's own intent, written from the same context window that produced the change. There is no model anywhere in the storage or retrieval path, so the same query returns the same answer today and in two years, across model versions. Tools that infer reasoning post-hoc are running a second LLM that never saw the original prompt: what it produces is paraphrase, and re-running it can produce different categories for the same change. As a Hacker News commenter put it about a competing approach, "grep won't find your commit because you rejected 'oauth-library'… unless there is deterministic enforcement" (0x457).

Determinism alone is no longer a separator — OpenLore is deterministic-native too, and says so. The compound that separates is append-only testimony: reasoning the agent wrote itself, kept in a store where a rejection is a first-class record rather than an inactive status to be swept up.

Why "entity-level" matters. Most tools attribute lines. Selvedge attributes things you actually search for: users.email, env/STRIPE_SECRET_KEY, api/v1/checkout, deps/stripe. The first question after git blame is usually "what's the history of this column", not "what's the history of lines 40–48 of users.py".

Why "captured live" matters. Not a differentiator on its own — every tool here claims some flavour of it — but it's the mechanism that makes the reasoning trustworthy. Writing at the moment of the change, from the context that produced it, is the reason there's no second model in the path to hallucinate an explanation. An empty reasoning field is itself an honest signal: the agent didn't have one.

Comparison current as of 2026-08-05; OpenLore at v2.1.8 / 265★, verified against its source. Corrections welcome as an issue.

Why "changesets" matter. A Stripe billing rollout touches the users table, two new env vars, three new API routes, one dependency, and four functions across the codebase. Tag every event with changeset:add-stripe-billing and you can pull the entire scope back later — even if the original PR was broken into eight smaller ones over a month.

Selvedge ↔ Agent Trace. Agent Trace is an open AI code-attribution wire format published by Cursor (RFC, Jan 2026). Its original GitHub home went 404 in August 2026 and the multi-vendor momentum behind it has faded, but the spec and schema still resolve at agent-trace.dev, frozen at v0.1.0. Since v0.3.9, selvedge export --format agent-trace emits Agent Trace v0.1.0 records and selvedge import --format agent-trace reads them back — a portable, documented interchange format for file/line AI attribution, with reasoning and entity-level provenance carried in each record's dev.selvedge metadata. The mapping is in docs/agent-trace-interop.md; Selvedge vendors the schema and has no runtime dependency on the upstream project.

---

Quickstart

Claude Code — install the plugin (recommended)

Two commands, inside Claude Code. No prior pip install — the plugin bootstraps the server itself via uvx (or pipx):

/plugin marketplace add masondelan/selvedge
/plugin install selvedge@selvedge

That's the whole agent-facing surface in one step:

• the MCP server — 8 tools (log_change, prior_attempts, blame, diff, history, changeset, search, stale_decisions);
• a skill that tells the agent when to call them — before editing a tracked entity, after any substantive change;
• the PreToolUse enforcement hook — schema/migration edits are blocked until prior_attempts has been checked this session, with the prior reasoning in the block message;
• slash commands — /selvedge:status, /selvedge:blame <entity>, /selvedge:history, /selvedge:prior-attempts <entity>.

The store (.selvedge/selvedge.db) creates itself on the first logged change. Two optional extras stay CLI-side: the post-commit hook that stamps each event with its commit hash (selvedge install-hook), and — if you want the selvedge command on your own shell PATH — pip install selvedge, which the launcher then prefers over uvx for an exact pinned version.

Plugin or selvedge setup for Claude Code? Pick one. Both wire the MCP server; running both registers it twice. The plugin is the lighter path and the one that updates itself. If you're on the plugin and only want the post-commit commit-hash stamping, run selvedge install-hook on its own.

Choose your coding agent

With uv installed:

uv tool install --upgrade selvedge
selvedge demo
cd your-project
selvedge setup --agent codex

Use codex, claude-code, cursor, copilot, gemini or windsurf. Repeat --agent for multiple tools, or omit it to detect installed agents. Prefer pip? Use python -m pip install --upgrade selvedge in a virtual environment. The selvedge-server executable must be on your editor's PATH; launch the editor from that environment or use the executable's absolute path in its MCP config.

Setup asks before changing files, backs up existing content, installs MCP and agent instructions, initializes the project, and offers a Git post-commit hook. For Codex it writes .codex/config.toml and AGENTS.md; Gemini CLI gets .gemini/settings.json and GEMINI.md; Copilot gets .vscode/mcp.json and .github/copilot-instructions.md. Custom Codex TOML entries require manual reconciliation, even with --force.

Restart your agent in the project and approve Selvedge's tools if prompted. Codex must trust the project to load project-scoped configuration. Ask the agent:

Use Selvedge to record one approach we considered and rejected in this project. Include why, and what would change our mind. Then look it up with prior_attempts.

Start a new session and look up the same entity to verify that the decision carries forward. MCP access does not automatically capture every decision: the installed instructions guide the agent to use it. Session delivery and enforcement hooks are currently Claude Code integrations.

For CI bootstrap or devcontainer.json postCreateCommand:

selvedge setup --non-interactive --yes

Verify the wiring — open a second terminal in the same project:

selvedge watch

Make any change in your AI tool — add a column, rename a function, add an env var. selvedge watch should print the new event within a second of the agent calling log_change. If nothing arrives, run selvedge doctor for a single-command health check that tells you which step is silently broken.

Query your history:

selvedge status                        # recent activity + missing-commit count
selvedge diff users                    # all changes to the users table
selvedge diff users.email              # changes to a specific column
selvedge blame payments.amount         # what changed last and why
selvedge history --since 30d           # last 30 days of changes
selvedge history --since 15m           # last 15 minutes ('m' = minutes)
selvedge changeset add-stripe-billing  # all events for a feature/task
selvedge search "stripe"               # full-text search
selvedge stats                         # log_change coverage report (per-agent)
selvedge import migrations/            # backfill from migration files
selvedge export --format csv           # dump history to CSV

If you don't want to run the wizard, the four manual steps it automates:

1. Initialize in your project

cd your-project
selvedge init

2. Register the MCP server

Selvedge is a standard stdio MCP server, so it works with any MCP client — Claude Code, Cursor, Windsurf, Codex CLI, Gemini CLI, and more. See Works with any MCP client for the exact config per client. For Claude Code:

claude mcp add selvedge -- selvedge-server

3. Tell your agent to use it

selvedge prompt --install CLAUDE.md

Point --install at whichever prompt file your client reads — the block itself is identical across clients:

Client	Prompt file
Claude Code	CLAUDE.md
Codex CLI (and other AGENTS.md-aware tools)	AGENTS.md
Cursor	.cursor/rules/selvedge.md (or legacy .cursorrules)
Gemini CLI	GEMINI.md

This installs the canonical agent-instructions block, sentinel-bracketed (<!-- selvedge:start --> / <!-- selvedge:end -->) so future --install calls update the bracketed region without disturbing anything else in the file. Or pipe it:

selvedge prompt | tee -a CLAUDE.md

Prefer to copy-paste? The same block is one click away on the website: selvedge.sh/prompt-block — with a copy button and notes on what your agent does with it.

4. Install the post-commit hook

selvedge install-hook

That's the same four steps the wizard runs.

---

Works with any MCP client

Selvedge is a standard stdio MCP server — its launch command is selvedge-server, put on your PATH by pip install selvedge. Any MCP-capable client can run it. Pick yours:

claude mcp add selvedge -- selvedge-server

Or commit a project-level .mcp.json so your whole team gets it:

{
  "mcpServers": {
    "selvedge": { "command": "selvedge-server" }
  }
}

Docs: https://code.claude.com/docs/en/mcp

.cursor/mcp.json (project) or ~/.cursor/mcp.json (global):

{
  "mcpServers": {
    "selvedge": { "command": "selvedge-server" }
  }
}

Cursor's newer schema also accepts an explicit "type": "stdio"; the command-only form works too (Cursor infers stdio from command). Docs: https://cursor.com/docs/mcp

~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "selvedge": { "command": "selvedge-server" }
  }
}

Windsurf hot-reloads the file — no restart needed. The in-app Plugins → View raw config button opens the exact file Cascade reads. Docs: https://docs.windsurf.com/windsurf/cascade/mcp

~/.codex/config.toml:

[mcp_servers.selvedge]
command = "selvedge-server"

Or run codex mcp add selvedge -- selvedge-server. Docs: https://developers.openai.com/codex/config-reference

~/.gemini/settings.json (or .gemini/settings.json per project):

{
  "mcpServers": {
    "selvedge": { "command": "selvedge-server" }
  }
}

Or run gemini mcp add -s user selvedge selvedge-server. Docs: https://github.com/google-gemini/gemini-cli/blob/main/docs/tools/mcp-server.md

Most clients share the same JSON shape — point yours at:

{
  "mcpServers": {
    "selvedge": { "command": "selvedge-server" }
  }
}

If selvedge-server isn't found, use its absolute path (which selvedge-server).

---

How it works

Selvedge runs as an MCP server. AI agents in tools like Claude Code call Selvedge's tools as they work — logging structured change events to a local SQLite database.

Each event records:

• What changed (entity path, change type, diff)
• When (timestamp)
• Who (agent, session ID)
• Why (reasoning — captured from the agent's context in the moment)
• Where (git commit, project)

The diff is git's job. The why is Selvedge's.

---

Selvedge tracks its own history

This repo dogfoods Selvedge: its .selvedge/selvedge.db is committed, so a fresh clone ships with Selvedge's own why-history. Clone it and ask why any part of Selvedge changed:

git clone https://github.com/masondelan/selvedge
cd selvedge
selvedge status                       # recent changes to Selvedge itself
selvedge search "telemetry"           # why the opt-in heartbeat shipped
selvedge blame selvedge/semantic.py   # why semantic search was added

Every event was logged by the agents that built Selvedge — the same log_change calls this README asks you to make in your own project.

---

Entity path conventions

users.email           DB column (table.column)
users                 DB table
src/auth.py::login    Function in a file (path::symbol)
src/auth.py           File
api/v1/users          API route
deps/stripe           Dependency
env/STRIPE_SECRET_KEY Environment variable

Prefix queries work everywhere: users returns users, users.email, users.created_at, and any other entity under the users. namespace.

---

MCP tools

When connected as an MCP server, Selvedge exposes:

Tool	Description
log_change	Record a change event with entity, diff, and reasoning. rename_from + change_type="rename" records the dual-event rename pattern; change_type="supersede" re-opens a reverted decision (append-only); optional constraint / stale_when keep the decision's principle and its invalidation condition queryable
diff	History for an entity or entity prefix, each row annotated with superseded_by
blame	Most recent change + context for an exact entity, plus the derived decision status (active / reverted / reopened)
history	Filtered history across all entities
changeset	All events grouped under a named feature/task slug
search	Full-text search across all events
prior_attempts	Prior change attempts on an entity + inferred outcome (tried → reverted → re-opened) — call it before editing. Optional fuzzy query adds semantically similar records (needs the semantic extra; falls back to substring)
stale_decisions	Decisions due for a revisit: past their revisit_after and still in active use (flag="revisit_due"), or whose stale_when condition matched a later change (flag="review_suggested")

---

CLI reference