Compare commits

...

5 Commits

Author SHA1 Message Date
didericis-codex 9c19ce4c58 docs: redesign forge identity trust boundary
tracker-policy-pr / check-pr (pull_request) Successful in 4s
2026-07-26 18:32:07 +00:00
didericis-claude 010e253d66 docs: sharpen the attribution guarantee to a byte<->activation-key binding
tracker-policy-pr / check-pr (pull_request) Successful in 6s
Revised per PR #480 (#5607 owner clarification + #5608 codex resolution;
#5612 directs the update):

- The audit row no longer implies upstream observation or agent-only
  authorship. Reworded the guarantee: the row cryptographically binds
  commit bytes (control-plane-RECOMPUTED SHA) to access to the activation
  signing key, and binds that key to control-plane-owned activation
  metadata. An agent can sign arbitrary contents but cannot verify as a
  different activation or choose the recorded metadata.
- Control plane accepts gateway-delivered opaque bytes, independently
  recomputes the Git object ID, verifies the embedded signature against
  the activation key, and stamps its own metadata. Trusts no gateway
  SHA/key/verdict/metadata. No upstream fetch.
- Purged overclaims: removed "a compromised gateway cannot fabricate an
  audit binding" (the sidecar holds the signing capability, so it can —
  and that's acceptable under the intended guarantee), plus "accepted
  push" / "introduced upstream" framing.
- Resolved the control-plane-transport open question in-PRD (was left
  open; codex asked to resolve): transport is gateway bytes +
  recompute + verify; mirror-read is no stronger.

Issue: #423

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-26 02:23:13 +00:00
didericis-claude 245f258f20 docs: drop author/committer enforcement; anchor audit in control plane
tracker-policy-pr / check-pr (pull_request) Successful in 7s
lint / lint (push) Successful in 2m44s
Revised per PR #480 review (#5590 + didericis-codex review on d8362ec):

- Remove author/committer enforcement entirely (#5590). The gate no
  longer matches identity fields; author/committer are recorded as
  claims in the audit store. Drop the git-gate.signing.enforce knob
  (which also resolves codex issue 1: a knob that weakened the stated
  guarantee). Add a "Deferred: identity enforcement" section noting it
  as a possible future add. Rename PRD/file to "signed commits & audit
  attribution" since identity is no longer guaranteed.
- Fix control-plane vs data-plane verification (codex issue 2, PRD 0070):
  git-gate (data plane) does a synchronous pre-forward SIGNATURE check
  only; the orchestrator/control plane (sole owner of bot-bottle.db)
  independently re-verifies each signature before writing attributed_commit.
  A gateway assertion alone never creates an audit row. New "Trust
  boundary" + "Control-plane verification & recording" sections.
- Reframe the guarantee to signed provenance + host-owned, independently
  verified audit record; ADR 0002 "claimed, not vouched" posture kept.
- attributed_commit now records claimed author/committer columns.

Issue: #423

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-25 22:18:48 -04:00
didericis-claude c62d57d5ac docs: narrow PRD to per-bottle signed identity & audit attribution
Revised per PR #480 review (#5518#5556):

- Rename: "forge subroles" → "per-bottle signed identity & audit
  attribution"; rename the file to match.
- Reframe the guarantee as bottle/activation provenance, not
  cryptographically-vouched author identity. Author/committer name/email
  is a claim carried inside the signed object, made trustworthy by a
  git-gate acceptance check + the host record, not by the signature.
- Add the gate-side acceptance check: on push, every newly-introduced
  commit (excluding upstream-reachable history) must verify against the
  activation key AND match git-gate.user in both author and committer
  fields, else the push is rejected. Host verifies the signature before
  recording a SHA as attributed.
- Audit: retain full public key + fingerprint + principal + validity
  interval (not fingerprint-only); state allowed-signers generation.
- Drop from scope: forge subuser accounts, provisioned API tokens/PAT
  minting, forge status/Verified badges -> future "forge actors" PRD.
  This removes the Gitea PAT bootstrap problem entirely.
- Manifest: drop git-forge/forge-accounts; reuse git-gate.user as the
  enforced identity + add opt-in git-gate.signing. Push stays PRD 0048
  deploy keys, unchanged.

Issue: #423

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-25 22:18:48 -04:00
didericis-claude 34bb7263fa docs: PRD for forge subroles (per-bottle subuser identity & vouched attribution)
Formalizes the design settled in issue #423: one forge subrole identity
per bottled agent (author + forge account + SSH signing key), reused
across all repos/forges. Vouched attribution via sign-at-commit-time in
the git-gate boundary (forwarded ssh-agent; private key never in the
bottle; no SHA divergence). Forge "Verified" badges abandoned in favor
of local git verify-commit plus durable console audit records and
commit-status badges. Reprovision-per-activation credential lifecycle
(0048 discipline), fail-loud teardown, public-key-fingerprint-only audit
trail on bottled_agent.

Successor to PRD 0027 (claimed-not-vouched, ADR 0002) and PRD 0048
(host-side minting lifecycle).

Issue: #423

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-25 22:18:48 -04:00
@@ -0,0 +1,651 @@
# PRD prd-new: Trusted agent forge identity, signed commits & audit attribution
- **Status:** Draft
- **Author:** didericis-claude
- **Created:** 2026-07-25
- **Issue:** #423
## Summary
Give each **host-trusted agent definition** an author identity and named forge
accounts, and let a bottle associate each git-gate repository with one of those
accounts. The resolved association provisions authenticated forge API access
through the egress proxy and injects provider-specific workflow instructions
into the agent's system prompt without exposing the token. Repository-local
agent and bottle definitions are no longer discovered because they would
otherwise be able to select host credential references or impersonate an agent.
For bottles that opt into signing, give each bottled agent a **per-activation
signing key** so every commit it produces is signed in the git-gate trust
boundary (outside the bottle) and recorded in bot-bottle's **host-owned audit
store**, which is the portable source of truth. Each row cryptographically binds
a commit's bytes (and their control-plane-recomputed SHA) to **access to that
activation's signing key**, and binds the key to control-plane-owned activation
metadata — bottle, host, manifest, agent, activation interval, retained public
key, configured agent author — and separately records the commit's *claimed*
author/committer. The gate mints a short-lived Ed25519 key at spin-up, holds the
private half in the sidecar `ssh-agent`, and forwards only `SSH_AUTH_SOCK` into
the bottle. The gate rejects any commit it forwards that is not signed by the
activation key; separately, the **control plane** independently recomputes each
commit's object ID and verifies its signature before recording attribution — it
never trusts a SHA, key, or verdict asserted by the gate.
This PRD deliberately does **not** enforce or cryptographically vouch
author/committer identity. The trusted agent definition supplies the configured
name/email, while the values in each commit remain claims carried inside the
signed object; making the gate reject a mismatch is a possible future add (see
**Non-goals** and **Deferred: identity enforcement**). Git push capability stays
exactly as PRD 0048 deploy keys. Forge API identity uses an operator-provided,
agent-specific token referenced from the trusted host environment; bot-bottle
does not mint forge users or tokens.
Successor to:
- **PRD 0027 (agent git identity, #94)** / **ADR 0002** — established that
agent name/email is *claimed, not vouched*. This PRD moves that identity from
the bottle/git-gate overlay to the trusted agent definition and adds signed
**provenance** plus a durable host record, not identity enforcement.
- **PRD 0011 (per-file manifests)** — allowed repository-local agent files to
override home agents. This PRD removes that trust path: agents and bottles are
loaded only from the host-owned `~/.bot-bottle` tree.
- **PRD 0048 (deploy-key provisioning, #169)** — the host-side mint-at-spin-up /
revoke-at-teardown lifecycle the signing key follows. Deploy keys are
unchanged.
- **PRD 0070 (per-host orchestrator, #351)** — the orchestrator/control plane is
the sole owner of `bot-bottle.db`; audit verification and recording live
there, not in the data-plane gate (see **Trust boundary**).
## The guarantee
The crisp property this feature provides:
> The **host-owned audit store** binds a set of commit bytes — whose Git object
> ID the control plane **recomputes** itself — to **access to this activation's
> signing key**, and binds that key to **control-plane-owned activation
> metadata**: bottle, host, manifest, agent, activation interval, retained public
> key, configured agent author. An agent may author and sign arbitrary commit
> contents, but it cannot make that signature verify as a *different*
> activation, and it cannot choose the activation metadata the control plane
> records. The commit's author/committer identity is **recorded separately as a
> claim**, not enforced or vouched. The forge is an external transport and
> collaboration surface, not the source of attribution truth.
What this does and does not prove (issue #423, comments #5554 / #5607 / #5608):
- It proves **access to activation *Y*'s signing key**: whoever assembled these
commit bytes could sign with that key. Recomputing the object ID and verifying
the embedded signature binds the SHA to activation *Y*, and the control plane's
own records bind *Y*'s key to *Y*'s metadata.
- It does **not** prove the commit was ever pushed, observed upstream, kept
(vs. later reverted or dropped), or produced by the *agent* rather than by any
other holder of the activation signing capability (the sidecar itself). The
store deliberately makes no claim about publication or sole-agent authorship
— the owner's requirement is attribution of *what manifest/agent/etc. was in
use when a commit was signed*, not proof of where the commit went (#5607).
- It does **not** make author/committer identity cryptographically vouched. The
bottle chooses every byte sent through the forwarded agent, so a signature over
`author Mallory <mallory@example>` is just as valid. Those fields are a claim
carried inside the signed object and recorded as-is.
- The binding is trustworthy because the **control plane** supplies the SHA (it
recomputes it), the public key, and the activation metadata from its own state
— never from a value the gateway asserts (see **Trust boundary**). The
residual, by design: anything that holds the activation signing capability can
produce commits that attribute to that activation. That is inherent to a
binding on *activation-key access*, not a defect.
## Problem
An agent runs on the developer's machine as a *subrole*, scoped down per role.
Locally that is fine because the machine is single-tenant. The git history an
agent produces, however, is a durable artifact that outlives the session and
can be pushed to shared repositories, and today bot-bottle offers no
tamper-evidence over it:
- **No provenance.** Nothing ties a pushed commit to the bottle/activation that
actually produced it. The configured name/email is forgeable and cosmetic
(ADR 0002); a commit could be produced anywhere.
- **No durable, portable record.** There is no host-side ledger that says "SHA
*X* was produced by agent *A* in bottle *B* on host *H* during interval
*[t0,t1]*, signed by key *K*," independent of any forge and surviving key
rotation.
- **Forge workflow context is missing.** The agent prompt does not know which
forge backs a git-gate repository, which API base URL to use, or that
authenticated requests must go through the egress proxy. Bespoke prompt text
has drifted between agents, causing incorrect PR creation behavior such as
using Gitea AGit review refs instead of a branch-backed pull request.
- **Identity is owned by the wrong layer.** `git-gate.user` puts an agent
property on a repository transport component. The same author identity and
forge actor should follow the agent across bottles and repositories.
- **Repository-local agents are a trust escalation.** Today
`$CWD/.bot-bottle/agents/*.md` can override a host agent. Once an agent
definition may reference a forge token, allowing the checked-out repository
to choose that definition would let untrusted workspace content select host
identities and credentials.
## Goals / Success Criteria
- **Per-activation signing key.** A fresh Ed25519 keypair is minted host-side at
each activation; the private half lives only in the sidecar `ssh-agent`, never
in the bottle. Only `SSH_AUTH_SOCK` crosses the boundary.
- **Agent-owned identity.** Author name/email and named forge accounts live on
the agent definition, not under `git-gate`.
- **Bottle-owned repository policy.** Signing remains an opt-in property of the
bottle, and each bottle repository may associate itself with one named forge
account from the selected agent.
- **Forge-aware prompting.** The resolved agent+bottle manifest contributes a
generated, provider-specific system-prompt section describing the forge API
URL, proxy-authenticated access path, repository mapping, and safe PR
workflow. No token value or token environment-variable name appears in the
prompt.
- **Proxy-held forge credential.** The host resolves the forge account's token
reference and gives it only to the egress proxy, which injects authentication
for the configured forge origin. The bottle receives neither the token nor a
credential file containing it.
- **Trusted definitions only.** Agent and bottle files are discovered only
under the host-owned `~/.bot-bottle/{agents,bottles}` directories.
`$CWD/.bot-bottle/{agents,bottles}` never contributes definitions or
overrides.
- **Signed commits with no SHA divergence.** Commits produced in the bottle are
signed at commit time; the SHA the agent observes is the SHA that reaches the
upstream through the gate.
- **Gate rejects unsigned commits.** Before the gate forwards a push, every
newly-introduced commit (those not already reachable from the advertised
upstream refs) must verify against the activation public key; a push with any
unsigned or wrong-key new commit is rejected, loudly, with the offending SHA.
This is a **signature** check only — no author/committer matching.
- **Control-plane-owned attribution.** The orchestrator/control plane (sole
owner of `bot-bottle.db`, PRD 0070) recomputes each commit's object ID from the
bytes, verifies the embedded signature against the activation public key it
holds, and attaches activation metadata from its own state — accepting no SHA,
key, verdict, or metadata asserted by the gateway. No upstream fetch is
required.
- **Host is the source of truth.** The audit record binds each recomputed SHA to
the bottle, host, manifest, agent, configured agent author, activation
interval, and retained public key, and separately records the commit's claimed
author/committer.
- **Verifiable after teardown.** The audit record retains the **full public
key, fingerprint, principal, and validity interval** — enough to regenerate an
allowed-signers file and run `git verify-commit` long after the activation
ends and the key is gone.
- **Reprovision-per-activation, fail-loud teardown.** The signing key is minted
once per activation (persists across restarts within that activation) and
discarded at teardown; deploy-key revocation continues to follow PRD 0048's
fail-loud discipline.
- **Push capability unchanged.** Git transport remains PRD 0048 deploy keys.
The forge account token is for forge API actions such as opening and
commenting on pull requests; it is not used for Git push.
## Non-goals
- **Author/committer enforcement.** Explicitly out of scope for this PRD (issue
#423, comment #5590). The gate does not reject a commit for carrying a foreign
author or committer; those fields are recorded as claims. We rely on the
cross-forge audit store of signed commits and the authors recorded there. See
**Deferred: identity enforcement** for what a future add would look like.
- **Cryptographically-vouched author identity.** Not claimed — see **The
guarantee**.
- **Forge account or token minting.** Bot-bottle does not create subusers or
PATs. The operator creates the agent-specific account/token out of band and
names the host environment secret in the trusted agent definition.
- **Forge-side attribution surfaces.** No commit-status badges, no forge
"Verified" badge. The latter is doubly unsuitable: it renders dynamically
against a *currently registered* key (so it would lie the moment a
reprovisioned key is revoked), and on Gitea registering a signing key also
grants push. Attribution lives in the host record and local `git
verify-commit`, not the forge.
- **Non-Gitea forges, dashboard UI for orphan cleanup, mid-session rotation,
dirty-teardown reconciliation.** As before; a separate cleanup/sync pass
handles orphans left by a crash or discarded snapshot.
## Scope evolution
This PRD started as "forge subroles" (forge subuser accounts + provisioned API
tokens + optional forge status posting + signing). Review first narrowed it,
then restored only the part needed for correct agent operation:
1. **Dropped forge account/token provisioning** (#5518#5556): the PAT
bootstrap is not implementable as sketched (Basic-Auth-as-target-user
constraint), and the signature never vouched the author anyway. The host
audit store remains the portable source of truth.
2. **Dropped author/committer enforcement** (#5590): rely on the audit store of
signed commits and the authors recorded there; gate enforcement of the
identity fields is a possible future add, not part of this slice.
3. **Restored declarative forge accounts, not provisioning** (#6002): agents
still need an operator-supplied API identity and forge-specific system
instructions to open and update PRs correctly. The trusted agent definition
therefore references an existing host secret; bot-bottle neither creates nor
rotates that credential.
4. **Moved identity to the agent trust domain** (#6002): author and forge
accounts are agent properties. Bottle repositories select an account by
name, while git-gate remains transport-only. Because these references grant
access to host credentials, repository-local agent/bottle discovery is
removed.
The resulting feature is **trusted agent identity + operator-provided forge API
access + forge-aware prompting + signed commits + a host-owned,
independently-verified audit record.** Identity enforcement (the gate rejecting
a foreign author/committer) remains a candidate future PRD.
## Design
### Trust boundary (control plane vs data plane)
git-gate is the **data plane**: it parses hostile bytes from inside the bottle
and forwards pushes. The orchestrator is the **control plane** and, per PRD
0070, is the sole owner of `bot-bottle.db`. These are different trust boundaries,
and the audit record must be anchored in the control plane:
- The gate performs a **synchronous pre-forward signature check** (below) and
can reject a push before it reaches the upstream. This is a data-plane gate on
what leaves the bottle, not the audit binding.
- The **control plane** takes the commit bytes to attribute (gateway-delivered
opaque bytes are fine), **recomputes the Git object ID**, **verifies the
embedded signature** against the activation public key it minted and holds, and
writes `attributed_commit` attaching metadata from its own state. It accepts
**no** gateway-supplied `verified` flag, claimed SHA, public key, or activation
identity.
The precise trust statement (issue #423, review by didericis-codex on d8362ec,
resolved in #5608): the row binds *these commit bytes / this recomputed SHA* to
*access to this activation's signing key*, and the control plane binds that key
to the recorded activation metadata. It does **not** assert forge observation or
that only the agent (not the signing sidecar) authored the commit — so this PRD
does **not** claim a compromised gateway cannot obtain an attribution row.
Because the sidecar holds the activation signing capability, a compromised
gateway *can* assemble and sign a commit and have it attributed to that
activation; what it cannot do is make the signature verify as a *different*
activation or choose the metadata the control plane records. That residual is
acceptable under the intended guarantee (#5607) and is why the guarantee is
worded as activation-key access, not agent-only authorship or upstream
publication. The gate therefore cannot stand in for host-side verification: the
control plane recomputes the object ID and verifies the signature itself rather
than trusting the gate's word.
### Identity model
Per **bottled agent** (agent definition ∘ sealed bottle), realized per
activation:
| Part | Value | Source | Role |
|------|-------|--------|------|
| Signing key | one Ed25519 keypair | minted host-side per activation | signs every commit; private half sidecar-only; the anchor of provenance |
| Author/committer | name + email | trusted agent definition `author` | written into commits and **recorded** as a claim; **not** enforced |
| Forge actor | named account + API origin + token reference | trusted agent definition `forge-accounts` | authenticates forge API actions through the egress proxy |
| Repository/forge association | forge account name | bottle `git-gate.repos.<repo>.forge` | selects the actor and generated workflow guidance for that repository |
### Manifest surface
Identity belongs to the agent. Repository capability and policy belong to the
bottle. The following files are both host-owned:
```yaml
# ~/.bot-bottle/agents/claude.md
---
author:
name: didericis-claude
email: eric+claude@dideric.is
forge-accounts:
didericis-gitea:
auth:
type: token
token_secret: GITEA_CLAUDE_TOKEN
url: https://gitea.dideric.is/api/v1
---
```
```yaml
# ~/.bot-bottle/bottles/dev.md
---
git-gate:
signing:
enabled: true # NEW — opt-in per-activation signing + audit
repos:
bot-bottle:
url: ssh://git@100.78.141.42:30009/didericis/bot-bottle.git
provisioned_key: # PRD 0048 — push capability, UNCHANGED
provider: gitea
token_env: GITEA_DEPLOY_TOKEN
host_key: "ssh-ed25519 AAAA..."
forge: didericis-gitea # account from the selected agent definition
---
```
- `author` is agent-only and replaces the `git-gate.user` agent/bottle overlay.
It is required when `git-gate.signing.enabled` is true or when any selected
repository has a `forge` association. The resolved values populate
`user.name` and `user.email`.
- `forge-accounts` is an agent-only map keyed by the existing manifest
kebab-case identifier grammar (`[a-z][a-z0-9-]*`). Each entry contains:
- `url`: an HTTPS forge API base URL. This PRD supports Gitea API URLs; a
future provider must add explicit typed validation and prompt generation
rather than accepting arbitrary prompt text supplied by a repository.
- `auth.type`: `token` in this slice.
- `auth.token_secret`: the name of a host environment variable containing the
operator-provided, agent-specific API token. The value is resolved only by
host provisioning and passed only to the egress proxy.
- `git-gate.repos.<name>.forge` is bottle-only and must resolve to an account in
the selected agent definition. An unknown account, an unsupported forge URL,
a non-HTTPS URL, or a missing/empty host secret fails launch before creating
the bottle.
- `git-gate.signing.enabled: true` opts a bottle in. Without it, behavior is
exactly as today. There is **no `enforce` sub-key** — this PRD does not enforce
identity fields, so no knob is needed (and a knob that weakened a guarantee
was flagged as a contradiction in review).
- `git-gate.signing`, `git-gate.repos`, and their `forge` associations are
bottle-only. `author` and `forge-accounts` are agent-only. Validation errors
point to the correct file/type instead of silently ignoring misplaced keys.
- `provisioned_key.token_env` remains the deploy-key administration credential
from PRD 0048. It is separate from the forge actor's `token_secret`: the
former provisions Git push capability, while the latter performs API actions
as the agent.
- Existing `git-gate.user` fields fail with migration guidance to move the
values into the selected home agent's `author` block. There is no period where
bottle identity silently overrides agent identity.
### Definition trust and discovery
Only the host-owned manifest tree is authoritative:
- Agents: `~/.bot-bottle/agents/*.md`
- Bottles: `~/.bot-bottle/bottles/*.md`
`$CWD/.bot-bottle/agents/*.md` no longer contributes new agents and no longer
overrides a home agent. `$CWD/.bot-bottle/bottles/*.md` remains unusable. If
either repository-local directory contains manifest files, bot-bottle emits a
warning that they are ignored and points to the home paths. Agent enumeration,
`require_agent`, lazy loading, and dashboard selectors all use the same
home-only index so there is no alternate path that can still select a workspace
definition.
This intentionally supersedes PRD 0011's repository-agent overlay. Workspace
instructions remain repository content (for example `AGENTS.md`), but executable
runtime policy, host secret references, and actor identity do not.
Programmatic in-memory manifests remain available for tests and internal
composition; they are already supplied by trusted host code and are not a
filesystem discovery path.
### Forge API provisioning and generated prompt
For each distinct forge account referenced by the selected bottle's repos, the
host:
1. Parses and canonicalizes the HTTPS API origin and rejects credentials in the
URL, fragments, and unsupported path shapes.
2. Resolves `auth.token_secret` from the host environment. The secret value is
copied only into the egress proxy's credential environment.
3. Adds an inspected egress route scoped to that forge origin/API prefix with
the provider's authentication scheme (`token` for Gitea). Authentication is
injected by the proxy; the bottle sends an unauthenticated request to the
configured URL.
4. Appends a generated, non-secret section to bot-bottle's existing system
prompt file. The section is derived from validated typed fields, not copied
Markdown from a repository.
For the example above, the generated guidance communicates:
- account alias `didericis-gitea` and API base
`https://gitea.dideric.is/api/v1`;
- repository `bot-bottle` uses that account;
- forge API calls must use the configured HTTPS URL through the proxy and must
not read, print, or manually attach an authorization token;
- Git pushes still use the bottle's git-gate remote;
- create/update a normal `refs/heads/<branch>` and open a branch-backed Gitea
pull request through the API; do not push `refs/for/*`, `refs/draft/*`, or
`refs/for-review/*`;
- use the API for review/comment operations and verify the returned object/state
before claiming the action completed.
The prompt includes neither the token value nor `GITEA_CLAUDE_TOKEN`. Keeping
even the environment-variable name out of the bottle reduces accidental
credential probing and prevents bespoke agent prompts from needing secret
implementation details.
### Signing: sign at commit time via a forwarded ssh-agent
The reason SHAs never diverge:
- The **sidecar** (the git-gate trust boundary) runs an `ssh-agent` holding the
short-lived signing private key.
- **Only `SSH_AUTH_SOCK`** is forwarded into the bottle — a bounded signing
capability, not the key.
- The provisioner writes the bottle `.gitconfig`:
```ini
[commit]
gpgsign = true
[gpg]
format = ssh
[user]
name = didericis-claude
email = eric+claude@dideric.is
signingkey = ssh-ed25519 AAAA... # activation signing PUBLIC key
```
- `git commit` asks the forwarded agent to sign; the signature is embedded at
object creation, so the agent-space SHA equals the pushed SHA. No transcoder,
no SHA translation table.
### Gate pre-forward signature check (data plane)
The gate already fetches from upstream before every `upload-pack` and mirrors
bidirectionally (PRD 0008). When `git-gate.signing.enabled` is set, after
gitleaks and before forwarding a push upstream:
1. **Compute the newly-introduced set.** Commits reachable from the pushed ref
tips but **not** reachable from any ref already advertised by the upstream
(which the gate knows because it fetches upstream first) — equivalent to
`git rev-list <new-tips> --not <all-known-upstream-refs>`. This excludes
pulled/merged existing history; a merge commit the bottle creates is itself
new and is checked, its already-upstream ancestors are not.
2. **Verify each new commit's signature** against the activation public key. A
commit that is unsigned or signed by any other key causes the push to be
**rejected** with the offending SHA.
3. No author/committer matching is performed.
This is a synchronous safety gate on what leaves the bottle; it is not the audit
record.
### Control-plane verification & recording
For each commit to attribute (the gate hands the control plane the commit bytes;
opaque gateway-delivered bytes are acceptable because nothing the gateway *says*
about them is trusted), the orchestrator/control plane:
1. **Recomputes the Git object ID** from the bytes itself. The stored `sha` is
this recomputed value, never a SHA the gateway claims.
2. **Verifies the embedded signature** against the activation public key it
minted and holds for that activation (via a generated allowed-signers file) —
ignoring any `verified` flag, key, or activation identity supplied by the
gateway.
3. Writes `attributed_commit` only for bytes that pass, stamping the activation
metadata (bottle/manifest/agent/host/interval) from its **own** state — not
from anything the gateway provides — and recording the commit's claimed
author/committer.
No upstream fetch is required: the guarantee is a byte↔activation-key binding, so
the object does not need to come from the forge (issue #423, #5608). Bytes that
do not verify against the activation key are **not** recorded as attributed (they
may be logged as an anomaly instead).
### Audit trail
The host SQLite store (PRD 0067, `~/.bot-bottle/bot-bottle.db`, owned by the
control plane per PRD 0070) records the signing-key lifecycle and per-commit
attribution. Retention is the **full public key, fingerprint, principal, and
validity interval** — enough to regenerate an allowed-signers file and verify
commits after teardown (issue #423, comment #5554, resolution 3). Never any
private key material.
```sql
CREATE TABLE bottled_agent_activation (
bottled_agent_slug TEXT NOT NULL,
activation_id TEXT NOT NULL, -- one per activation cycle
host TEXT NOT NULL,
manifest_digest TEXT NOT NULL, -- ties the record to the sealed manifest
agent TEXT NOT NULL,
configured_author_name TEXT NOT NULL, -- trusted agent definition value
configured_author_email TEXT NOT NULL, -- trusted agent definition value
signing_pubkey TEXT NOT NULL, -- full ssh-ed25519 public key (for verify-commit)
signing_fpr TEXT NOT NULL, -- SHA256:... fingerprint (stable handle)
principal TEXT NOT NULL, -- allowed-signers principal, e.g. the author email
valid_from TEXT NOT NULL,
valid_until TEXT, -- NULL while active; set at teardown
status TEXT NOT NULL, -- active | retired
PRIMARY KEY (bottled_agent_slug, activation_id)
);
CREATE TABLE attributed_commit (
sha TEXT NOT NULL, -- control-plane-RECOMPUTED object ID, not gateway-claimed
bottled_agent_slug TEXT NOT NULL,
activation_id TEXT NOT NULL,
repo TEXT NOT NULL,
author_name TEXT NOT NULL, -- CLAIMED, recorded as-is (not enforced)
author_email TEXT NOT NULL, -- CLAIMED
committer_name TEXT NOT NULL, -- CLAIMED
committer_email TEXT NOT NULL, -- CLAIMED
observed_at TEXT NOT NULL,
PRIMARY KEY (sha, repo)
);
```
Verification/allowed-signers generation is a stated part of the design: for a
given SHA, join `attributed_commit → bottled_agent_activation`, emit
`<principal> <signing_pubkey>` to a temporary allowed-signers file, and
`git verify-commit` (or `ssh-keygen -Y verify`) against it. The
`(pubkey, principal, valid_from/until)` tuple is exactly what that requires. The
activation's `configured_author_*` columns preserve the trusted agent
configuration. The attributed commit's author/committer columns are the
commit's *claim*; consumers compare the two if useful, understanding that a
mismatch is recorded but not rejected.
### Credential lifecycle
Signing follows PRD 0048's lifecycle discipline. The operator-provided forge
actor token is referenced, not provisioned:
- **Activation:** mint a fresh Ed25519 signing keypair; load the private half
into the sidecar `ssh-agent`; write the public half into `.gitconfig` and the
`bottled_agent_activation` row (`active`, `valid_from` set). Deploy keys are
provisioned exactly as PRD 0048. Resolve each referenced forge actor token
from the host environment and install it only in the egress proxy process.
Minting is **per activation** (a restart re-attaches the same key; a new
activation mints a new key and retires the old row), so frozen snapshots
don't accumulate live keys.
- **Teardown (fail-loud):** revoke provisioned deploy keys via the forge API
(0048); discard the signing key from the sidecar agent and set the activation
row to `retired` with `valid_until`. The signing key was never on the forge,
so there is nothing to revoke there — only the local retire. Deploy-key
revocation failure halts teardown (0048); 404 = already-gone = success. Stop
the egress proxy to discard its copy of the forge actor token. The
operator-owned token itself is not revoked because bot-bottle did not mint it
and it may be reused by later activations of the same trusted agent.
- **Dirty teardown** is assumed handled; a separate cleanup/sync pass reconciles
orphaned deploy keys.
## Deferred: identity enforcement
If a future PRD wants the gate to *enforce* that new commits carry the manifest
identity, the natural shape is: extend the gate pre-forward check to also require
each new commit's author **and** committer name/email to equal the resolved
agent `author`,
rejecting mismatches — with the same control-plane re-verification before
recording. This is deliberately left out now (issue #423, comment #5590); it is
noted so the door stays open and the current schema (which records the claimed
author/committer) already carries what such a check would compare against. Note
that even then the property would be gate-*enforced*, not signature-*vouched*; a
validating signing broker in front of the key would be required for the latter.
## Implementation chunks
1. **This PRD.** Sets the revised design and trust boundary.
2. **Trusted definition boundary.** Remove `$CWD/.bot-bottle/agents` from
discovery, override, enumeration, lazy loading, and selectors. Keep both
agent and bottle definitions home-only; warn on ignored repository files.
Update PRD 0011-facing docs and migration guidance.
3. **Identity and forge manifest surface.** Add agent-only `author` and
`forge-accounts`; remove `git-gate.user`; add bottle-only
`git-gate.signing` (`enabled` only) and
`git-gate.repos.<name>.forge`. Validate account references after composing
the selected agent+bottle and fail closed on missing host secrets or
unsupported URLs/providers.
4. **Forge proxy + prompt provisioning.** Resolve referenced actor tokens into
scoped egress proxy routes and generate provider-specific, non-secret system
instructions from typed manifest fields. Gitea guidance covers API usage,
branch-backed PRs, prohibited AGit refs, and verifying mutations. Test that
neither token values nor token secret names enter the bottle or prompt.
5. **Signing pipeline.** Sidecar `ssh-agent` provisioning; forward
`SSH_AUTH_SOCK` into the bottle across docker, smolmachines, macOS-container,
and firecracker backends; emit the `commit.gpgsign` / `gpg.format=ssh` /
`user.signingkey` gitconfig. Integration test: a bottle commit is
`verify-commit`-valid and its SHA is unchanged through the gate; the private
key is absent from the bottle.
6. **Gate pre-forward signature check.** Compute the newly-introduced set
(excluding upstream-reachable commits), verify each against the activation
key, reject unsigned/wrong-key with the offending SHA. Tests: unsigned
rejected; wrong-key rejected; pulled/merged upstream history passes; an
all-signed push succeeds. A foreign-author commit that is correctly signed
**passes the gate** (identity is not enforced here).
7. **Control-plane verification + audit.** `bottled_agent_activation` /
`attributed_commit` tables (PRD 0067 store, control-plane-owned per PRD 0070);
the control plane recomputes each commit's object ID and verifies the
signature before writing a row; retain full pubkey + fingerprint + principal +
validity interval; record claimed author/committer; allowed-signers generation
+ a post-teardown `verify-commit` helper. Tests: a gateway-claimed SHA/key/
verdict is ignored — the row's `sha` is the recomputed ID and bytes not signed
by the activation key produce **no** row.
8. **Docs.** Glossary entries for forge account and per-bottle signed commits;
README agent/bottle schema and home-only migration; ADR note that
signing-enabled bottles gain signed *provenance* and a host-owned audit
record while authorship stays *claimed* (ADR 0002 unchanged).
## Testing strategy
- **Unit — trust boundary (must):** cwd agent files are ignored with a warning,
cannot override a home agent, are absent from enumeration/selectors, and
cannot be loaded by name. Cwd bottle behavior remains home-only.
- **Unit — manifest (must):** `author`, `forge-accounts`,
`git-gate.signing`, and repo `forge` parsing/validation; misplaced-field
rejection; kebab-case account names; unknown account references; HTTPS/API
URL validation; missing token-secret environment values.
- **Unit — prompt/proxy (must):** only referenced forge accounts produce proxy
routes and guidance; Gitea instructions name the API/repo and branch-backed
workflow; token values and `token_secret` names are absent from the prompt and
bottle environment; auth is scoped to the validated forge API origin.
- **Integration — signing (must):** end-to-end signed commit verifies with
`git verify-commit`; SHA observed in the bottle equals the SHA upstream; the
private key is absent from the bottle.
- **Integration — gate check (must):** unsigned rejected; wrong-key rejected;
the upstream-reachable exclusion (pull + merge human history and push a signed
merge); a correctly-signed foreign-author commit **passes** (no identity
enforcement); a clean all-signed push succeeds.
- **Control plane (must):** the control plane recomputes the object ID and
records a row for bytes genuinely signed by the activation key; a gateway-
supplied SHA/key/verdict is ignored (the stored `sha` is the recomputed value);
bytes signed by a foreign/invalid key produce **no** row; the activation
retains the configured agent author while a differing commit author is
recorded separately as an unenforced claim.
- **Lifecycle:** activation mints the key and writes an `active` row; teardown
retires it (`valid_until`) and revokes deploy keys fail-loud; a restart
re-attaches the same key (no new row); a fresh activation mints a new key and
retires the old.
- **Post-teardown verification:** regenerate the allowed-signers file from a
`retired` row and confirm `verify-commit` still succeeds for an attributed SHA.
## Resolved: control-plane transport
Raised in review and **resolved** (issue #423, #5608): the commit object does not
need to come from the forge, and reading the gateway-owned mirror is no stronger
than accepting gateway-delivered bytes — both are fabricatable, and neither
matters because the control plane trusts nothing the gateway *asserts*. The
transport is therefore: the gate hands the control plane the raw commit bytes,
the control plane **recomputes the object ID** and **verifies the signature**
against the activation key, and stamps its own activation metadata. No upstream
fetch. This is exactly what makes the byte↔activation-key binding sound
regardless of transport.
## Open questions
- **Where the gate check slots into PRD 0008 ordering.** Modeled as a
pre-forward step after gitleaks; confirm it composes with the existing
access-hook / mirror ordering rather than needing a separate hook.