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, Cursor, Copilot) 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 v0.3.10

The memory comes to the agent, and the store gets its dials. Two themes, shipped together because the config half is what the rest needed to read settings from.

Delivery. Selvedge already blocked re-edits of reverted entities. What was missing was delivery when there is nothing to veto. Two new hooks:

• SessionStart injects a compact digest as a session begins — decisions due for a revisit, entities that were tried and reverted, recent changesets.
• PreCompact fires just before context compaction destroys this session's reasoning and names the watched entities you edited but never logged.

Both are quiet when they have nothing to say, size-capped, read-only, and templated. Neither can block anything — PreCompact deliberately declines the veto the hook API offers it. This is the answer to a measured failure mode: two 2026 papers recorded pull-model memory tools going unused entirely (zero voluntary memory operations across 114 turns against a pre-seeded store) while deterministic injection landed every time.

selvedge export --format markdown renders the store as a reviewable digest to commit next to it, so captured intent shows up in a pull request instead of hiding inside a binary. Deterministic — regenerating with no new events is a zero-line diff.

Config. .selvedge/config.toml is now first-class, with a canonical precedence chain that selvedge doctor prints per setting. It brings:

• selvedge prune --include-events — the first path that can delete captured reasoning, so it needs both a confirmation and SELVEDGE_DESTRUCTIVE=1. Neither alone is enough, because --yes in a cron entry defeats a prompt and a shell profile defeats an env var. Events retention defaults to never.
• Event-size bounds (diff_bytes, reasoning_bytes) that truncate loudly — a marker in the text, a warning at write time, a count in selvedge stats.
• Secret-shape warnings at log_change, extendable via redaction_patterns, plus a doctor row that scans what's already stored. Warn, never reject.

Also: five review issues closed. The enforcement hook's allow path is 40% faster (33.6 ms → 20.1 ms per gated call) and SELVEDGE_HOOK_DISABLE=1 finally short-circuits before the imports it was documented to skip; log_change no longer discards revisit_after / constraint / stale_when on renames and supersedes; the CLI's --json and the MCP tools now return identical structures; and the Docker image no longer ships the maintainer's own database. Tests 826 → 984.

What's new in v0.3.9.3

Fixes a broken install, and lands a full code-quality pass. mcp 2.0.0 (released 2026-07-28) removed mcp.server.fastmcp, and Selvedge declared mcp>=1.0.0 with no upper bound — so any pip install selvedge after that date pulled 2.0.0 and selvedge-server failed at import. This release pins the dependency. If your server stopped starting, this is why — upgrade.

It ships alongside a review that put nine independent passes over the codebase and then tried to disprove every finding before acting on it. Seventeen confirmed defects fixed. The ones you would actually have noticed:

• The enforcement hook stopped blocking things it shouldn't. Reading a tracked file — cat, git diff, pytest, ruff check — was blocked, and the remediation the error message told you to run was blocked by the same gate, so there was no way out from the CLI. Two more paths fed the same false blocks: a commented-out line of SQL counted as a real deletion, and any commit message merely containing the word "revert" marked every file it touched as reverted.
• Lookups got fast at scale. The main entity read was scanning every row — measured 7.4 ms → 0.35 ms at 100k events, and the hook had been taking seconds on large stores.
• selvedge setup can no longer delete parts of your CLAUDE.md, an interrupted backup can no longer destroy your last good one, and upgrading while two Selvedge processes are running no longer crashes with an error that looked like database corruption.

Tests went 739 → 826. No schema change and no tool-surface change, so this is drop-in for anyone on 0.3.9.x.

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.

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 (Cursor + Cognition AI, RFC Jan 2026, backed by Cloudflare, Vercel, Google Jules, Amp, OpenCode, and git-ai) is an emerging open standard for AI code attribution traces. Selvedge isn't a competitor to it — it's a compatible producer. As of v0.3.9, selvedge export --format agent-trace emits Agent Trace v0.1.0 records (and selvedge import --format agent-trace reads them back); the mapping is in docs/agent-trace-interop.md. Agent Trace is the wire format. Selvedge is the live capture + query layer that emits it.

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.

Any other MCP client — selvedge setup

Cursor, Copilot, Windsurf, Codex CLI, Gemini CLI, and the rest:

pip install selvedge
cd your-project
selvedge setup

That's it. selvedge setup is an interactive wizard: it detects which AI tools you have (Claude Code, Cursor, Copilot), writes the MCP entry into each one's config, drops the canonical agent-instructions block into your project's prompt file (CLAUDE.md / .cursorrules / copilot-instructions.md), installs the PreToolUse enforcement hook into .claude/settings.json (Claude Code only — blocks schema/migration edits until prior_attempts has been checked; --skip-enforcement-hook to opt out), runs selvedge init, and installs the post-commit hook. Every modified file gets a .bak written next to it before any change reaches disk. Re-running is a no-op.

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

cd your-project
selvedge init

claude mcp add selvedge -- selvedge-server

selvedge prompt --install CLAUDE.md

selvedge prompt | tee -a CLAUDE.md

selvedge install-hook

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

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

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

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

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

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

{
  "mcpServers": {
    "selvedge": { "command": "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:

CLI reference

selvedge init [--path PATH]               Initialize in project
selvedge status                           Recent activity summary
selvedge diff ENTITY [--limit N]          Change history for entity
selvedge blame ENTITY                     Most recent change + context
selvedge history [--since SINCE]          Browse all history
              [--entity ENTITY]
              [--project PROJECT]
              [--changeset CS]
              [--summarize]
              [--limit N]
selvedge changeset [CHANGESET_ID]         Show events in a changeset
                  [--list]                or list all changesets
                  [--project NAME]
                  [--since SINCE]
selvedge search QUERY [--limit N]         Full-text search
selvedge prior-attempts ENTITY            Prior attempts + inferred outcome,
                       [--description T]   with the tried → reverted →
                       [--all]             re-opened trail + status line
                       [--window 7d]       (--all widens recall)
                       [--fuzzy TEXT]      add semantic matches (needs the
                                           semantic extra; substring fallback)
selvedge supersede ENTITY                 Re-open a reverted decision —
                  --reasoning TEXT         append-only, links the prior
                  [--constraint TEXT]      reverted event (or --supersedes ID)
                  [--stale-when TEXT]
                  [--supersedes ID]
selvedge index [--model NAME]             Build/update the optional semantic
              [--json]                     embeddings index (selvedge[semantic])
selvedge stale [--entity ENTITY]          Decisions due for a revisit: past
              [--project NAME]            revisit_after + still in use, or
              [--agent NAME]              stale_when matched by a later change
              [--json]                    ("review suggested")
selvedge stats [--since SINCE]            Tool call coverage report (per-tool, per-agent)
selvedge doctor [--json]                  Health check: DB path, schema, hook, MCP wiring
selvedge install-hook [--path PATH]       Install git post-commit hook
                     [--window MIN]       (default 60 minutes)
selvedge backfill-commit --hash HASH      Backfill git_commit on recent events
                        [--window MIN]    (default 60 minutes)
selvedge import PATH                      Import migrations (SQL / Alembic) or
              [--format auto|sql|         an Agent Trace file (agent-trace)
                 alembic|agent-trace]
              [--from-git]                or walk git history for reverts:
              [--since REF|DATE]          revert-message commits + deletions
              [--project NAME]            become change_type="revert" events
              [--dry-run]                 (idempotent on commit + entity)
selvedge export [--format json|csv|       Export history (agent-trace =
                 markdown|agent-trace]      Agent Trace v0.1.0 records;
                                            markdown = reviewable digest)
              [--since SINCE]
              [--entity ENTITY]
              [--ndjson]                  agent-trace: one record per line
              [--collapse-by-session]     agent-trace: merge a session into one
              [--output FILE]
selvedge log ENTITY CHANGE_TYPE           Manually log a change
             [--diff TEXT]                CHANGE_TYPE: add, remove, modify,
             [--reasoning TEXT]           rename, retype, create, delete,
             [--agent NAME]               index_add, index_remove, migrate,
             [--commit HASH]              revert, supersede
             [--project NAME]
             [--changeset CS]
             [--revisit-after WHEN]       ISO date or offset (e.g. 90d)
             [--rename-from OLD]          OLD path when CHANGE_TYPE is 'rename'
             [--constraint TEXT]          the principle behind the decision
             [--stale-when TEXT]          what would invalidate it
             [--supersedes ID]            with CHANGE_TYPE 'supersede'
selvedge migrate-paths                    Re-canonicalize stored entity paths
                      [--apply]           (dry-run by default; --apply writes)
                      [--json]

All read commands support --json for machine-readable output.

Relative time in --since:

• 15m → last 15 minutes (m = minutes)
• 24h → last 24 hours
• 7d → last 7 days
• 5mo → last 5 months (mo or mon = months)
• 1y → last year

Unparseable inputs (e.g. --since yesterday) exit with a clear error rather than silently returning empty results. ISO 8601 timestamps are also accepted and normalized to UTC.

Configuration

.selvedge/config.toml

Every key is optional; a missing file means the defaults below. Precedence is CLI flag → env var → project .selvedge/config.toml → global ~/.selvedge/config.toml → default. SELVEDGE_DB is the one exception: it always wins for database resolution, because the config file is found by resolving that path. selvedge doctor prints the effective value and the step that produced it for every setting.

retention_days_events     = 0       # 0 = never delete events (the default)
retention_days_tool_calls = 90      # local telemetry retention
backup_keep_last          = 7
diff_bytes                = 65536   # truncate oversized diffs at log time
reasoning_bytes           = 32768   # truncate oversized reasoning
db_size_warn_mb           = 500     # doctor warns above this
stale_days                = 0       # 0 = off
digest_max_bytes          = 4096    # cap on the session-start digest
redaction_patterns        = []      # extra secret shapes to warn about

[hook]
watch_globs = ["**/migrations/**", "db/**/*.sql"]

Every key also has an env override (SELVEDGE_DIFF_BYTES, SELVEDGE_RETENTION_DAYS_EVENTS, …).

Reviewing captured intent in a pull request

.selvedge/selvedge.db is a SQLite file, so the reasoning inside it doesn't show up in a diff. Export a Markdown digest next to it and commit both:

selvedge export --format markdown -o .selvedge/DECISIONS.md
git add .selvedge/

The digest is grouped by entity with reverted decisions first, and it is deterministic — regenerating with no new events produces a zero-line diff, so it stays reviewable instead of becoming noise everyone learns to skip. Heading anchors derive from the entity path, so links into it keep working as it grows. Regenerate it in the same commit as the code, or from a pre-commit hook.

Coverage checking

Wondering how often your agent actually calls log_change? Two ways to check:

# Quick summary in the terminal
selvedge stats

# Cross-reference against git commits
python scripts/coverage_check.py --since 30d

The coverage script compares your git log against Selvedge events and shows which commits have associated change events. Low coverage usually means the system prompt needs strengthening — see docs/fallbacks.md for guidance.

In CI (GitHub Action)

The same check ships as the Selvedge Coverage Check composite Action, so you can track agent coverage on every push — and optionally fail the build when it drops:

# .github/workflows/selvedge-coverage.yml
name: Selvedge coverage
on: [push, pull_request]
jobs:
  coverage:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0            # full history so commits can be matched
      - uses: masondelan/selvedge@v0.3.10   # pin to a release tag (or @main for latest)
        with:
          since: 30d
          fail-under: "0.5"         # optional: fail below 50% coverage; omit to report only

It writes a coverage summary to the job summary and exposes coverage-ratio, covered, and total as step outputs. The action cross-references your git history against the Selvedge event log, so the runner needs the project's .selvedge/selvedge.db (commit it, or restore it before this step) and full git history (fetch-depth: 0). Inputs: since, window, limit, fail-under, selvedge-version, python-version, working-directory, db-path.

Contributing

git clone https://github.com/masondelan/selvedge
cd selvedge
pip install -e ".[dev]"
pytest

See CLAUDE.md for architecture details and the phase roadmap.

License

MIT — see LICENSE.