Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions docs/agents/ACCESS.md
Original file line number Diff line number Diff line change
Expand Up @@ -15,8 +15,9 @@ it as the revocation checklist. Written as OSTIF-audit follow-up T-14 (G-012).

| Privilege | Holder | Notes |
|---|---|---|
| GitHub org owner (`git-agentic`) | Toni Bergholm | Org 2FA requirement **enforced** since 2026-07-18 (see [`opsec-baseline.md`](../security/opsec-baseline.md)). |
| Repo admin (`git-agentic/src-control`) | Toni Bergholm | Controls rulesets, security toggles, secrets. |
| GitHub org owner (`git-agentic`) | Toni Bergholm (`tonibergholm`) | Org 2FA requirement **enforced** since 2026-07-18 (see [`opsec-baseline.md`](../security/opsec-baseline.md)). |
| Repo admin (`git-agentic/src-control`) | `tonibergholm`, `tonibergholm-codento` | Both admin. Control rulesets, security toggles, secrets. |
| Reviewer (required PR approver) | `tonibergholm`, `tonibergholm-codento` | Since 2026-07-18 the `main` ruleset requires **1 approval** from a principal other than the last pusher, so the two accounts review each other's PRs (closes audit G-032). |
| `git-agentic.com` DNS/registrar | Toni Bergholm | Out-of-repo credential. |
| crates.io publish rights | — (N/A) | Workspace is `publish = false`; no crates.io tokens exist. Revisit at first release (audit T-24 trigger). |
| Security-report handler | Toni Bergholm | Per [`vulnerability-response.md`](../security/vulnerability-response.md). |
Expand Down
24 changes: 24 additions & 0 deletions docs/audit/gap-analysis-source.md
Original file line number Diff line number Diff line change
Expand Up @@ -668,3 +668,27 @@ Two closing items are human/out-of-repo, not code: deploy `security.txt` to
feed subscriptions listed in `docs/security/subscriptions.md` (G-024 tail).
No gap was withdrawn; the gap set stands at 35 and the report's Section 0
carries the remediation status.

## Fifth-round delta: required reviews enabled (2026-07-18)

Later the same day the maintainer added a second admin account
(`tonibergholm-codento`) and tightened the `main` ruleset. Verified live via
`gh api repos/git-agentic/src-control/rulesets/18739705`: the `pull_request`
rule now has `required_approving_review_count: 1`,
`dismiss_stale_reviews_on_push: true`, and `require_last_push_approval: true`.

- **GAP-032/T-16 is now fully closed.** It was the last non-deferred gap and
had been held only because a lone maintainer cannot approve their own PR; a
second reviewing principal now exists and approvals are required repo-wide.
- **Scorecard consequence:** `CodeReviewID` (#14) and `BranchProtectionID`
(#4) will improve/close on the next weekly scan (approvals now recorded;
the branch-protection warnings for require-approvers, stale-dismissal, and
last-push-approval are all now satisfied — `require_code_owner_review`
stays false, a minor residual). The only remaining open code-scanning
finding is the optional `CIIBestPracticesID` badge (#16).
- **Operational note:** no principal — human, automation, or agent — can
self-merge anymore; the two admin accounts review each other's PRs.

Every non-deferred gap in the 35-gap set is now resolved. Remaining: the
Deferred tier T-22…T-25 (gated on a first distribution channel) and the two
human/out-of-repo tails (`security.txt` deploy, subscription confirmation).
16 changes: 10 additions & 6 deletions docs/audit/ostif-best-practices-audit.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@ title: "Gap Analysis: src-control vs OSTIF Security Best Practices Guide"
comparison_direction: "src-control repository (current state) -> OSTIF/Least Authority Security Best Practices Guide (desired state)"
scope: "The whole repository (code, CI, docs, GitHub configuration via live API) compared against all six OSTIF practice areas; the guide's missing 08-what-next.md chapter and the external git-agentic.com website were excluded."
generated: "2026-07-18"
updated: "2026-07-18 — fourth round: Now + Soon tiers remediated (PRs #76–#87); code-scanning down from 17 open alerts to 3, all tracing to one deferred item (required human approvals) plus an optional badge"
updated: "2026-07-18 — fifth round: second maintainer + required PR approvals enabled, closing G-032/T-16; code-scanning now 1 open alert (the optional OpenSSF badge). Every non-deferred gap is resolved."
generated_by: "han:gap-analysis"
sections_included:
- executive_summary
Expand Down Expand Up @@ -40,9 +40,9 @@ Every gap has a stable ID. Sections 3–5 reference those IDs. The full evidence

> This section was added after the gap analysis was acted on. Sections 1–5 below describe the repository **as first audited**; this section records what has since been fixed. Where they differ, this section is current.

**Both actionable tiers of the work plan are done.** The Now tier (T-1…T-9, T-26) and the Soon tier (T-10…T-21) all shipped across pull requests #76–#87, plus repository-settings changes and an organization-wide two-factor-authentication requirement. Only the **Deferred tier (T-22…T-25)** remains — OSS-Fuzz enrollment, reproducible builds, a release pipeline with supply-chain attestations, and a hardening guide for people who run the server component — and every one of those is deliberately parked behind a trigger that hasn't happened yet: the project still distributes nothing for a consumer to verify.
**Both actionable tiers of the work plan are done, and the one item that had been deferred for lack of a second reviewer is now also closed.** The Now tier (T-1…T-9, T-26) and the Soon tier (T-10…T-21) all shipped across pull requests #76–#87, plus repository-settings changes, an organization-wide two-factor-authentication requirement, and — as of later on 2026-07-18 — a second maintainer and required pull-request approvals (closing G-032/T-16). Only the **Deferred tier (T-22…T-25)** remains — OSS-Fuzz enrollment, reproducible builds, a release pipeline with supply-chain attestations, and a hardening guide for people who run the server component — and every one of those is deliberately parked behind a trigger that hasn't happened yet: the project still distributes nothing for a consumer to verify.

**The independent security scorecard tells the story in numbers: 17 open findings became 3.**
**The independent security scorecard tells the story in numbers: 17 open findings became 3, and after required reviews were enabled only 1 remains — the optional badge.**

| Scorecard finding (first run) | Now/Soon fix | Status |
|---|---|---|
Expand All @@ -55,10 +55,13 @@ Every gap has a stable ID. Sections 3–5 reference those IDs. The full evidence

The **automated code analysis is fully clean** — zero open findings — and now covers the desktop application's language as well as the core (T-3).

**The 3 remaining findings are not new gaps and not oversights — they are one already-tracked deferred item plus one optional extra:**
**Update (2026-07-18, later the same day): the two review-related findings are now resolved.** A second maintainer account was added, and the trunk protection was set to require **1 approval** from someone other than the last pusher, with stale approvals dismissed on push and last-push approval required. That closes gap G-032 / todo T-16 — it was the only remaining non-deferred gap — and the two Scorecard findings it drove (required review approvals, and branch-protection completeness) clear on the next weekly scan. Nobody self-merges now; the two admin accounts review each other's pull requests.

- **Two findings (required review approvals, and branch-protection completeness) share a single root cause: review approvals are set to zero.** This is gap G-032 / todo T-16 in the plan, and it is deliberately held: under the hosting platform's rules a lone maintainer cannot approve their own change, so requiring approvals needs a second reviewer to exist first. Note the branch-protection finding already *improved* once merge-gating on passing checks landed (T-9); the rest of it unlocks together the day a second reviewer is added and approvals can be required.
- **One finding is an optional best-practices badge**, now showing "in progress." Completing it is nice-to-have, not a gap the guide requires.
**That leaves one open code-scanning finding, and it is optional:**

- **An optional best-practices badge**, now showing "in progress." Completing it is nice-to-have, not a gap the guide requires.

*(For the audit trail: these findings originally shared the root cause "review approvals set to zero," which was deliberately held because a lone maintainer cannot approve their own change under the platform's rules. Adding a second reviewer removed that blocker. The branch-protection finding had already improved once merge-gating on passing checks landed via T-9.)*

**Two closing items still need a human and cannot be done from the repository:** deploying the machine-readable security-contact file to the project website (`/.well-known/security.txt`), and the maintainer personally confirming the security-feed subscriptions the documentation now lists. Both are noted in Section 5.

Expand Down Expand Up @@ -717,6 +720,7 @@ The prioritized work plan, consolidated by the project-manager synthesis from th
5. **The two corrected over-credits** (dead private-reporting channel; CI runs-but-does-not-gate) are already reflected in Section 1 of this report.
6. **(Third round) The Scorecard `VulnerabilitiesID` alert (#18) will not close by fixing anything** — 17 of its 18 advisories are documented accepted risks and the 18th becomes one via T-26; its OSV check does not honor cargo-audit ignores, so expect to dismiss it with a comment pointing at `audit.yml`'s removal gates (or live with a standing 0-score on that check until the GTK4 migration).
7. **(Third round) Do not treat the same-day scorecard PRs (#74/#75) as closing T-1's intent** — the forcing function only works if someone actually triages what it surfaces; 17 alerts are open and this report's todo mapping is the triage.
8. **(Fifth round) Required approvals are now on — no self-merge.** Every PR needs 1 approval from a principal other than the last pusher, with last-push approval; the two admin accounts (`tonibergholm`, `tonibergholm-codento`) review each other's work. Automation and agents cannot merge their own PRs. This closed G-032/T-16.

---

Expand Down
7 changes: 5 additions & 2 deletions docs/security/opsec-baseline.md
Original file line number Diff line number Diff line change
Expand Up @@ -17,8 +17,11 @@ T-20 (G-025).
`checks: write` + `issues: write` to the job level).
- **Third-party actions are SHA-pinned** with a version comment; Dependabot
maintains the pins (audit T-2). New actions enter SHA-pinned or not at all.
- **Merges to `main` are ruleset-gated**: PR required, four required status
checks, up-to-date branch, no force-push/deletion, no bypass actors.
- **Merges to `main` are ruleset-gated**: PR required, **1 required approval**
from a principal other than the last pusher (since 2026-07-18), stale
approvals dismissed on push, last-push approval required, four required
status checks, up-to-date branch, no force-push/deletion, no bypass actors.
Nobody self-merges — the two admin accounts review each other's PRs.

## Accounts and devices

Expand Down
10 changes: 6 additions & 4 deletions docs/security/vulnerability-response.md
Original file line number Diff line number Diff line change
Expand Up @@ -38,10 +38,12 @@ class of work that must never enter that pipeline:

A security-fix PR requires **a human review distinct from the author** before
merge, regardless of who or what authored it. Agent-authored security fixes are
never merged on the strength of another agent's approval (audit G-032). For
other (non-security) changes the standing CI gates apply; this stricter rule is
scoped to security fixes until a second reviewing principal exists to enforce
it repo-wide.
never merged on the strength of another agent's approval (audit G-032). Since
2026-07-18 the `main` ruleset enforces this repo-wide for **all** PRs — 1
required approval from a principal other than the last pusher, with stale
approvals dismissed on push and last-push approval required — so a security fix
gets the same distinct-reviewer gate as any other change, plus the human-author
requirement above for agent-authored fixes.

## Steps on receipt

Expand Down