chore: update badge URLs to commit fe21232e [skip ci]

Co-authored-by: emo <emo@oblachno.com>

.devin/skills/dependency-graph/SKILL.md

@@ -0,0 +1,135 @@
+# dependency-graph
+
+Map of the oblachno ecosystem. Knows which repo produces what, which
+repos depend on which, and the correct order for cross-repo changes.
+
+## When to Invoke
+
+Invoke this skill when:
+- Changes span multiple repos
+- A change in one repo requires version bumps in downstream repos
+- Deploying infrastructure that depends on published packages/images
+- Verifying the ecosystem is in a consistent state before deployment
+- Determining which repos to update and in what order
+
+## Prerequisites
+
+- All repos cloned under `/home/emo/dev/ideas/oblachno/`
+- `.env` with `DEVELOPER_GITEA_API_TOKEN` in each repo
+
+## Ecosystem Map
+
+```
+                    devx (PyPI package)
+                   /    |    \
+                  /     |     \
+            grm   sso-bridge   infra
+           (PyPI)  (PyPI+Docker) (deploys all)
+              |        |           |
+              v        v           v
+         infra bump  infra bump   staging
+         (auto PR)   (auto PR)   production
+                                      |
+                          mattermost-oidc (Docker image)
+                          (infra pulls :latest at deploy)
+```
+
+## Repositories
+
+| Repo | Produces | Consumers | Release Trigger |
+|------|----------|-----------|-----------------|
+| `devx` | PyPI package `devx` | grm, sso-bridge, infra | User-facing changes to `src/devx/**` |
+| `grm` | PyPI package `grm` | infra | User-facing changes to `src/grm/**` or `ansible/**` |
+| `sso-bridge` | PyPI package `sso_bridge` + Docker image | infra | User-facing changes to `src/sso_bridge/**` or `ansible/**` |
+| `infra` | Staging/production deployment | (end users) | User-facing changes + nightly gate |
+| `mattermost-oidc` | Docker image `mattermost-oidc` | infra (pulls at deploy) | `Dockerfile` or `build.yml` changes |
+
+## Dependency Chain
+
+### devx → all repos
+
+devx publishes to the Gitea PyPI registry. grm, sso-bridge, and infra
+pin devx in `pyproject.toml`:
+```toml
+"devx @ git+https://git.oblachno.oblachno.fyi/oblachno-oss/devx.git@vX.Y.Z"
+```
+
+When devx publishes a new version:
+1. grm, sso-bridge, and infra must bump their pinned devx version
+2. This is currently manual — no auto-dependency-PR from devx
+3. Each repo must `make setup` to pick up the new version
+
+### grm → infra
+
+grm publishes to PyPI. Its post-merge workflow auto-creates an infra
+dependency PR via `devx.ci.create_dependency_pr --repo oblachno/infra
+--package grm`. The PR bumps the pinned grm version in infra's
+`pyproject.toml`.
+
+### sso-bridge → infra
+
+sso-bridge publishes to PyPI AND builds a Docker image. Its post-merge
+workflow auto-creates an infra dependency PR via
+`devx.ci.create_dependency_pr --repo oblachno/infra --package
+sso_bridge`. The PR bumps the pinned sso_bridge version.
+
+The Docker image is pulled by infra at deploy time (`sso-bridge:latest`).
+
+### mattermost-oidc → infra
+
+mattermost-oidc builds a Docker image tagged `:latest` and `:MM_VERSION`.
+infra pulls `mattermost-oidc:latest` at deploy time. There is no
+auto-dependency-PR — infra simply pulls the latest image.
+
+### infra → staging/production
+
+infra deploys to staging and production. The deployment:
+1. Provisions VMs from golden images
+2. Runs Ansible roles (including grm and sso-bridge roles)
+3. Pulls Docker images (sso-bridge, mattermost-oidc)
+4. Configures services
+
+## Correct Order for Cross-Repo Changes
+
+When a change spans multiple repos, follow this order:
+
+1. **devx first** — if the change starts in devx, merge and publish devx
+   first. Wait for the PyPI publish job to complete.
+2. **Bump devx in consumers** — in grm/sso-bridge/infra, bump the pinned
+   devx version, run `make setup`, verify tests pass, merge.
+3. **grm/sso-bridge second** — merge and publish grm/sso-bridge. Wait
+   for the PyPI publish + Docker image build to complete.
+4. **Auto-dependency-PRs** — grm/sso-bridge post-merge auto-creates infra
+   PRs to bump pinned versions. Wait for these PRs to appear.
+5. **Merge infra dependency PRs** — review and merge the auto-created
+   infra PRs.
+6. **infra last** — deploy to staging, validate, promote to production.
+
+## State Verification Before Deployment
+
+Before deploying infra, verify:
+
+1. **devx version consistent** — all repos pin the same devx version
+2. **grm published** — latest grm tag exists in PyPI
+3. **sso-bridge published** — latest sso_bridge tag exists in PyPI
+4. **sso-bridge image built** — latest sso-bridge Docker image exists
+5. **mattermost-oidc image built** — latest mattermost-oidc image exists
+6. **infra pins match published versions** — no stale pins
+7. **Nightly gate green** — `NIGHTLY_STATUS` is not `failed`
+
+## Quick Check Commands
+
+```bash
+# Check latest devx version
+curl -sS https://git.oblachno.oblachno.fyi/api/v1/repos/oblachno-oss/devx/releases/latest | python3 -c "import json,sys; print(json.load(sys.stdin).get('tag_name','?'))"
+
+# Check pinned devx version in each repo
+for repo in grm sso-bridge infra; do
+  echo -n "$repo: "; grep 'devx @' /home/emo/dev/ideas/oblachno/$repo/pyproject.toml | grep -oP 'v[\d.]+'
+done
+
+# Check latest sso-bridge image build
+curl -sS -H "Authorization: token $DEVELOPER_GITEA_API_TOKEN" \
+  "https://git.oblachno.oblachno.fyi/api/v1/repos/oblachno/sso-bridge/actions/runs?per_page=5" \
+  | python3 -c "import json,sys; [print(r['id'],r['status'],r['conclusion']) for r in json.load(sys.stdin).get('workflow_runs',[]) if r.get('event')=='push']"
+```

.devin/skills/deployment-coordination/SKILL.md

@@ -0,0 +1,78 @@
+# deployment-coordination
+
+How grm releases propagate to infra. grm publishes a PyPI package and
+auto-creates an infra dependency PR. Coordination ensures the PR is
+merged before infra deploys.
+
+## When to Invoke
+
+Invoke this skill when:
+- Changes to grm affect infra deployments
+- Preparing a grm release that infra depends on
+- Verifying infra has bumped to the latest grm version
+- Coordinating a multi-repo change that includes grm
+
+## Prerequisites
+
+- grm repo at `/home/emo/dev/ideas/oblachno/grm`
+- `.env` with `DEVELOPER_GITEA_API_TOKEN`
+- See `dependency-graph` skill for the full ecosystem map
+
+## What grm Produces
+
+grm publishes a Python package to the Gitea PyPI registry. infra pins it:
+```toml
+"grm @ git+https://git.oblachno.oblachno.fyi/oblachno-oss/grm.git@vX.Y.Z"
+```
+
+## Release Flow
+
+1. PR merged to master
+2. Post-merge workflow runs `devx.ci.release` — classifies changes
+3. If user-facing changes: git-cliff bumps version, creates tag, pushes
+4. `devx.ci.publish` builds and publishes to Gitea PyPI registry
+5. `devx.ci.create_dependency_pr --repo oblachno/infra --package grm`
+   auto-creates an infra PR to bump the pinned grm version
+
+## Downstream Consumer
+
+| Repo | Pin location | Auto-bump? |
+|------|-------------|------------|
+| infra | `pyproject.toml` | Yes — auto PR created by post-merge |
+
+## Coordinating a grm Change
+
+1. **Merge grm PR** — wait for post-merge publish + dependency-PR creation
+2. **Verify publish** — check the new tag:
+   ```bash
+   curl -sS -H "Authorization: token $DEVELOPER_GITEA_API_TOKEN" \
+     https://git.oblachno.oblachno.fyi/api/v1/repos/oblachno-oss/grm/releases/latest \
+     | python3 -c "import json,sys; print(json.load(sys.stdin).get('tag_name','?'))"
+   ```
+3. **Find the auto-created infra PR** — check infra for open PRs with
+   `dependency` label or title containing `bump grm`
+4. **Review and merge the infra dependency PR** — verify the version bump,
+   run `make pytest-cov` in infra, add `ready-to-merge`
+5. **Verify infra staging deploy** — after infra merges, staging deploy
+   picks up the new grm version
+
+## State Verification
+
+```bash
+# Current grm version
+grep '__version__' /home/emo/dev/ideas/oblachno/grm/src/grm/__init__.py
+
+# What infra pins
+grep 'grm @' /home/emo/dev/ideas/oblachno/infra/pyproject.toml | grep -oP 'v[\d.]+'
+
+# Check for open infra dependency PRs
+curl -sS -H "Authorization: token $DEVELOPER_GITEA_API_TOKEN" \
+  "https://git.oblachno.oblachno.fyi/api/v1/repos/oblachno/infra/pulls?state=open" \
+  | python3 -c "import json,sys; [print(p['number'],p['title']) for p in json.load(sys.stdin) if 'grm' in p.get('title','').lower()]"
+```
+
+## Common Mistakes
+
+- Merging the grm PR but ignoring the auto-created infra dependency PR
+- Deploying infra before the dependency PR is merged — stale grm version
+- Forgetting that grm also has an Ansible role used by infra at deploy time

.devin/skills/skill-creation/SKILL.md

@@ -0,0 +1,142 @@
+# skill-creation
+
+How to create, validate, and maintain Devin skills. Skills must be
+clear, succinct, and actionable — no AI slop.
+
+## When to Invoke
+
+Invoke this skill when:
+- Creating a new skill
+- Amending an existing skill
+- Evaluating whether a skill is needed
+- Reviewing a PR that adds or modifies skills
+
+## Prerequisites
+
+- Skill directory: `.devin/skills/<skill-name>/SKILL.md`
+- Validator: `tests/test_skills.py` (infra) or reference to it
+- Tests: `tests/unit/test_skills_validation.py` (infra)
+
+## When to Create a Skill
+
+Create a skill when:
+- An agent struggles with a task repeatedly (branch hygiene, PR order)
+- A workflow has non-obvious ordering constraints (deployment coordination)
+- A task requires specific tool usage over raw commands (CI monitoring)
+- Multiple agents need shared context (dependency graph)
+
+Do NOT create a skill for:
+- One-off tasks (use a spec instead)
+- Tasks already covered by AGENTS.md
+- Tasks that are obvious from the Makefile or README
+- Tasks that change frequently (skills should be stable)
+
+## Skill Structure
+
+Every skill MUST have:
+
+```markdown
+# <skill-name>
+
+One-line description of what the skill does.
+
+## When to Invoke
+
+2-4 bullet points describing when to use this skill.
+
+## Prerequisites
+
+What must exist before using the skill (venv, .env, tools).
+
+## <Core Content>
+
+The actual guidance. Keep it actionable.
+
+## Verification (if applicable)
+
+How to verify the skill's guidance works.
+
+## Common Mistakes (if applicable)
+
+What agents get wrong without this skill.
+```
+
+## Quality Standards
+
+### Do
+- **Be specific.** Reference exact make targets, file paths, commands.
+- **Be concise.** Each section should be scannable in under 30 seconds.
+- **Be actionable.** Every paragraph should tell the agent what to DO.
+- **Use tables** for command reference, mappings, and comparisons.
+- **Use code blocks** for commands the agent should run.
+- **Link to other skills** when related (e.g., "See `dependency-graph` skill").
+
+### Don't
+- **No preamble.** Don't start with "This skill helps agents..." — just state what it does.
+- **No filler.** Don't repeat information from AGENTS.md or other skills.
+- **No vague advice.** "Be careful with branches" is useless. "Run `git branch --show-current` before every commit" is useful.
+- **No AI slop.** Don't write "In this comprehensive guide, we will explore..." — just give the guidance.
+- **No redundant sections.** If "Common Mistakes" would repeat "When to Invoke", skip it.
+- **No marketing.** Don't describe the skill as "powerful" or "comprehensive".
+
+## Scope Rules
+
+- **One skill per concern.** Don't mix branch hygiene with CI monitoring.
+- **Project-specific, not generic.** Skills reference this repo's make targets, file paths, and conventions — not abstract advice.
+- **Shared skills must be identical across repos.** Use `SHARED_SKILLS` in the validator to enforce this.
+- **Per-repo skills must reflect that repo's reality.** Don't copy infra-specific targets to sso-bridge.
+
+## Automated Validation
+
+Every skill must pass the validator (`tests/test_skills.py`). The validator checks:
+
+1. **Structure** — H1 title, "When to Invoke" section, "Prerequisites" section
+2. **Commands** — referenced `make <target>` commands exist in Makefile or devx.mak
+3. **Paths** — referenced file paths exist in the repo
+4. **Shared skills** — identical content across repos (SHA-256 comparison)
+5. **No drift** — no references to nonexistent commands or files
+
+Run the validator:
+```bash
+python3 tests/test_skills.py --repo infra --repo sso-bridge
+```
+
+## Effectiveness Evaluation
+
+### Static Checks (automated, CI)
+
+The validator runs in CI as part of `make pytest-cov`. A failing skill
+test blocks the PR. This catches:
+- Missing sections
+- Invalid commands
+- Broken file references
+- Cross-repo drift
+
+### Runtime Metrics (manual, periodic)
+
+Track these signals to evaluate skill effectiveness:
+- **Skill invocation frequency** — how often agents invoke the skill
+- **Success rate when invoked** — did the skill prevent the mistake it targets?
+- **Feedback issues** — agents create Gitea issues with `feedback` label when a skill is unclear or wrong
+- **Mistake recurrence** — if agents still make the mistake the skill targets, the skill needs improvement
+
+### Retrospective Review
+
+Periodically (monthly or after major incidents) review skills:
+1. List all skills and their last-modified dates
+2. Check for feedback issues tagged `skill-improvement`
+3. Verify referenced commands still exist (run validator)
+4. Remove skills that are no longer relevant
+5. Update skills where mistakes still recur
+6. Document lessons in this skill's "Common Mistakes" section
+
+## Creating a New Skill — Checklist
+
+- [ ] Identify the repeated struggle or non-obvious workflow
+- [ ] Check no existing skill covers it
+- [ ] Write the skill following the structure above
+- [ ] Run `python3 tests/test_skills.py` — must pass
+- [ ] Run `make pytest-cov` — must pass with 100% coverage
+- [ ] If shared across repos, copy identical content to each repo
+- [ ] Add the skill to `SHARED_SKILLS` in the validator if shared
+- [ ] Create PR, verify CI passes, merge

CHANGELOG.md

@@ -2,6 +2,12 @@
 
 All notable changes to this project will be documented in this file.
 
+## [0.23.0] - 2026-08-28
+
+### Features
+
+- Add pre-cache timer, force_pull, and Docker socket options to runner config
+
 ## [0.22.1] - 2026-08-26
 
 ### Bug Fixes

README.md

@@ -8,12 +8,12 @@ Each runner runs in an isolated **rootless Docker** environment under a dedicate
 
 [![CI](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions/workflows/ci.yml/badge.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
 [![License: GPL-3.0](https://img.shields.io/badge/license-GPL--3.0-blue)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/src/branch/master/LICENSE)
-[![Coverage](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/coverage.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
+[![Coverage](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/coverage.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
-[![Tests](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/tests.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
+[![Tests](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/tests.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
-[![Docs](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/docs.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/wiki)
+[![Docs](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/docs.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/wiki)
-[![Code Quality](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/quality.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
+[![Code Quality](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/quality.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
-[![Version](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/version.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/releases)
+[![Version](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/version.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/releases)
-[![Python](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/python.svg)](https://www.python.org/downloads/)
+[![Python](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/python.svg)](https://www.python.org/downloads/)
 
 ## Why GRM?
 

ansible/roles/gitea_runner/defaults/main.yml

@@ -90,6 +90,22 @@ gitea_runner_remove_user: true
 gitea_runner_log_level: "info"
 gitea_runner_container_label: "gitea-runner=true"
 gitea_runner_file: ".runner"
+# force_pull: when false (default), the runner reuses locally cached images
+# instead of pulling on every job. Pre-cached images (via the pre-cache timer
+# or pre_pull_images task) eliminate registry thundering-herd when all runners
+# start jobs simultaneously.
+gitea_runner_force_pull: false
+# Container options passed to `docker run` for CI job containers.
+# Mounts the host rootless Docker socket as /run/host-docker.sock so
+# start_docker.py inside the container can detect and use the host daemon
+# (full disk, no nested DinD) instead of starting an inner dockerd.
+gitea_runner_container_options: "-v /run/user/{{ gitea_runner_uid }}/docker.sock:/run/host-docker.sock"
+# Volumes allowed in CI job containers (validated by the runner against
+# container.options and job-level volumes). Must include the host Docker
+# socket mount target.
+gitea_runner_valid_volumes:
+  - "/run/host-docker.sock"
+  - "/run/user/{{ gitea_runner_uid }}/docker.sock"
 
 # Containerd version pinning — Docker 28.x vendors containerd v2.1.x internally.
 # containerd.io >= 2.3 ships a shim that returns a protobuf BootstrapResult which
@@ -135,3 +151,11 @@ gitea_runner_docker_ipv6_cidr: "fd00:dead:beef::/48"
 # disk-space prune only removes dangling images, so pre-pulled tagged images persist.
 # Set to [] to skip pre-pulling. Images are pulled as the runner user via rootless Docker.
 gitea_runner_pre_pull_images: []
+
+# Pre-cache timer: periodically pulls the runner container image so it stays
+# fresh in the local Docker cache. This prevents thundering-herd registry
+# timeouts when all runners start CI jobs simultaneously with empty caches.
+# Runs every 6 hours (aligned with prune schedule). Set to empty string to
+# disable the timer.
+gitea_runner_pre_cache_schedule: "*-*-* 00/6:30:00"
+gitea_runner_pre_cache_images: "{{ gitea_runner_pre_pull_images }}"

ansible/roles/gitea_runner/tasks/main.yml

@@ -20,6 +20,9 @@
 - name: Include pre-pull images
   ansible.builtin.include_tasks: pre_pull_images.yml
 
+- name: Include pre-cache timer
+  ansible.builtin.include_tasks: pre_cache.yml
+
 - name: Include integration test
   ansible.builtin.include_tasks: integration_test.yml
   when: not gitea_runner_skip_registration

ansible/roles/gitea_runner/tasks/pre_cache.yml

@@ -0,0 +1,67 @@
+---
+# Periodic timer that pre-pulls CI runner images into the local Docker cache.
+# Prevents thundering-herd registry timeouts when all runners start jobs
+# simultaneously with empty/stale caches. Runs every 6 hours (configurable).
+# The prune timer removes dangling images but NOT tagged ones, so pre-pulled
+# images persist between runs.
+
+- name: Create docker-pull-images user service file
+  ansible.builtin.template:
+    src: docker-pull-images.service.j2
+    dest: "{{ gitea_runner_home }}/.config/systemd/user/docker-pull-images.service"
+    owner: "{{ gitea_runner_service_user }}"
+    group: "{{ gitea_runner_service_user }}"
+    mode: "0644"
+  register: gitea_runner_pre_cache_service
+
+- name: Create docker-pull-images user timer file
+  ansible.builtin.template:
+    src: docker-pull-images.timer.j2
+    dest: "{{ gitea_runner_home }}/.config/systemd/user/docker-pull-images.timer"
+    owner: "{{ gitea_runner_service_user }}"
+    group: "{{ gitea_runner_service_user }}"
+    mode: "0644"
+  register: gitea_runner_pre_cache_timer
+
+- name: Reload systemd user daemon for pre-cache timer
+  ansible.builtin.command: systemctl --user daemon-reload
+  become: true
+  become_user: "{{ gitea_runner_service_user }}"
+  environment:
+    XDG_RUNTIME_DIR: "/run/user/{{ gitea_runner_uid }}"
+    DBUS_SESSION_BUS_ADDRESS: "unix:path=/run/user/{{ gitea_runner_uid | default(0) }}/bus"
+  changed_when: true
+  when:
+    - gitea_runner_systemd_available.stat.exists
+    - gitea_runner_docker_rootless_setup
+    - gitea_runner_pre_cache_service is changed or gitea_runner_pre_cache_timer is changed
+    - gitea_runner_pre_cache_schedule | length > 0
+    - gitea_runner_pre_cache_images | length > 0
+
+- name: Enable and start docker-pull-images user timer
+  ansible.builtin.command: systemctl --user enable --now docker-pull-images.timer
+  become: true
+  become_user: "{{ gitea_runner_service_user }}"
+  environment:
+    XDG_RUNTIME_DIR: "/run/user/{{ gitea_runner_uid }}"
+    DBUS_SESSION_BUS_ADDRESS: "unix:path=/run/user/{{ gitea_runner_uid | default(0) }}/bus"
+  changed_when: true
+  when:
+    - gitea_runner_systemd_available.stat.exists
+    - gitea_runner_docker_rootless_setup
+    - gitea_runner_pre_cache_schedule | length > 0
+    - gitea_runner_pre_cache_images | length > 0
+
+- name: Disable and stop docker-pull-images timer (no images or schedule)
+  ansible.builtin.command: systemctl --user disable --now docker-pull-images.timer
+  become: true
+  become_user: "{{ gitea_runner_service_user }}"
+  environment:
+    XDG_RUNTIME_DIR: "/run/user/{{ gitea_runner_uid }}"
+    DBUS_SESSION_BUS_ADDRESS: "unix:path=/run/user/{{ gitea_runner_uid | default(0) }}/bus"
+  changed_when: true
+  failed_when: false
+  when:
+    - gitea_runner_systemd_available.stat.exists
+    - gitea_runner_docker_rootless_setup
+    - gitea_runner_pre_cache_schedule | length == 0 or gitea_runner_pre_cache_images | length == 0

ansible/roles/gitea_runner/tasks/pre_pull_images.yml

@@ -8,6 +8,15 @@
 #
 # Set gitea_runner_pre_pull_images to a list of image refs to pull, or
 # empty list to skip pre-pulling.
+#
+# IMPORTANT: Do NOT use this mechanism for:
+#   - CI runner container images (e.g. ci-full) — these are already
+#     cached by the runner setup task and pulling them here is redundant.
+#   - Images that molecule tests pull themselves — molecule prepare/converge
+#     steps handle their own image pulls; pre-pulling them here wastes time
+#     and disk space.
+# This mechanism is intended only for images that are needed by the runner
+# itself but not pulled by any molecule scenario or runner setup step.
 
 - name: Pre-pull Docker images for CI runner
   ansible.builtin.command: "docker pull {{ item }}"

ansible/roles/gitea_runner/templates/docker-pull-images.service.j2

@@ -0,0 +1,14 @@
+[Unit]
+Description=Pre-pull Docker images for CI runner cache
+After=docker.service
+Wants=docker.service
+
+[Service]
+Type=oneshot
+Environment=DOCKER_HOST=unix:///run/user/{{ gitea_runner_uid }}/docker.sock
+Environment=XDG_RUNTIME_DIR=/run/user/{{ gitea_runner_uid }}
+# Pull each image quietly. docker pull exits 0 if image is already up-to-date,
+# so this is idempotent. Errors are non-fatal (image may already be cached).
+{% for image in gitea_runner_pre_cache_images %}
+ExecStart=/usr/bin/docker pull -q {{ image }}
+{% endfor %}

ansible/roles/gitea_runner/templates/docker-pull-images.timer.j2

@@ -0,0 +1,10 @@
+[Unit]
+Description=Periodic Docker image pre-cache for CI runner
+
+[Timer]
+OnCalendar={{ gitea_runner_pre_cache_schedule }}
+Persistent=true
+RandomizedDelaySec=300
+
+[Install]
+WantedBy=timers.target

ansible/roles/gitea_runner/templates/gitea-runner-config.yaml.j2

@@ -9,3 +9,13 @@ runner:
 container:
   label: "{{ gitea_runner_container_label }}"
   docker_host: "unix:///run/user/{{ gitea_runner_uid }}/docker.sock"
+  force_pull: {{ gitea_runner_force_pull | lower }}
+{% if gitea_runner_container_options | length > 0 %}
+  options: "{{ gitea_runner_container_options }}"
+{% endif %}
+{% if gitea_runner_valid_volumes | length > 0 %}
+  valid_volumes:
+{% for volume in gitea_runner_valid_volumes %}
+    - "{{ volume }}"
+{% endfor %}
+{% endif %}

docs/index.md

@@ -8,12 +8,12 @@ Each runner runs in an isolated **rootless Docker** environment under a dedicate
 
 [![CI](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions/workflows/ci.yml/badge.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
 [![License: GPL-3.0](https://img.shields.io/badge/license-GPL--3.0-blue)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/src/branch/master/LICENSE)
-[![Coverage](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/coverage.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
+[![Coverage](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/coverage.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
-[![Tests](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/tests.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
+[![Tests](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/tests.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
-[![Docs](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/docs.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/wiki)
+[![Docs](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/docs.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/wiki)
-[![Code Quality](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/quality.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
+[![Code Quality](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/quality.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/actions)
-[![Version](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/version.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/releases)
+[![Version](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/version.svg)](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/releases)
-[![Python](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/5ba6b90dffc24876035267268dd3780f7089264c/python.svg)](https://www.python.org/downloads/)
+[![Python](https://git.oblachno.oblachno.fyi/oblachno-oss/grm/raw/commit/fe21232ef9dd554a49e4309bd918d358194b65d8/python.svg)](https://www.python.org/downloads/)
 
 ## Overview
 

docs/specs/GRM-162.md

@@ -0,0 +1,47 @@
+# GRM-162: Add pre-cache timer, force_pull, and Docker socket options to runner config
+
+## Problem
+CI containers were not using the host's rootless Docker daemon, leading to
+"no space left on device" errors. The runner config template was missing
+`force_pull`, `options` (host Docker socket mount), and `valid_volumes`
+fields. Additionally, no pre-cache timer existed to prevent thundering-herd
+registry timeouts when all runners pull images simultaneously.
+
+## Approach
+REQ-1: Add `force_pull: false` to runner config template (explicit default
+so the runner reuses locally cached images instead of pulling on every job)
+REQ-2: Add `options` field to mount host rootless Docker socket as
+`/run/host-docker.sock` so `start_docker.py` inside CI containers can detect
+and use the host daemon (full disk, no nested DinD)
+REQ-3: Add `valid_volumes` list for the socket mount targets (validated by
+the runner against `container.options` and job-level volumes)
+REQ-4: Add `pre_cache.yml` task with a systemd user timer that pre-pulls CI
+images every 6 hours (configurable via `gitea_runner_pre_cache_schedule`)
+REQ-5: Add `docker-pull-images.service.j2` and `docker-pull-images.timer.j2`
+templates for the pre-cache timer
+REQ-6: Timer is disabled when `gitea_runner_pre_cache_schedule` is empty or
+`gitea_runner_pre_cache_images` is empty (graceful degradation)
+
+## Test Plan
+- `make lint-ci` passes (ansible-lint on new task/template files)
+- `make molecule` converges successfully with the new pre-cache tasks
+- Verify the runner config template renders correctly with and without
+  container options/valid_volumes
+
+## Deploy Plan
+- Merge to master → post-merge auto-publishes package
+- Infra dependency PR auto-created to bump pinned grm version
+- Runners pick up the new config on next `make setup` or ansible apply
+
+## Rollback Plan
+- Revert the merge commit
+- Set `gitea_runner_pre_cache_schedule: ""` to disable the timer without
+  reverting
+
+## Acceptance Criteria
+- [x] REQ-1: `force_pull: false` in runner config template
+- [x] REQ-2: `options` field mounts host Docker socket as `/run/host-docker.sock`
+- [x] REQ-3: `valid_volumes` list includes both socket mount targets
+- [x] REQ-4: `pre_cache.yml` task creates and manages systemd user timer
+- [x] REQ-5: Service and timer templates created
+- [x] REQ-6: Timer disabled gracefully when schedule or images empty

docs/specs/GRM-172.md

@@ -0,0 +1,38 @@
+# GRM-172: Audit and document pre-pull image usage guidelines
+
+## Problem
+The grm repo contains a runner-level `pre_pull_images.yml` task file that
+pre-pulls Docker images to avoid repeated pulls on every CI run. However,
+there was no audit confirming that molecule `prepare.yml` files are not
+also redundantly pre-pulling images that the runner setup already caches.
+Wasteful pre-pulling wastes CI time and disk space.
+
+## Approach
+Audit all molecule `prepare.yml` files in the grm repo for pre-pull tasks.
+The audit found NO molecule prepare.yml files contain pre-pull tasks, so no
+code removal is needed. Document the audit findings in a spec and add a
+comment to the runner-level `pre_pull_images.yml` task file clarifying that
+it should not be used for images that molecule tests pull themselves (to
+avoid redundant pulls).
+
+REQ-1: Audit all molecule prepare.yml files for pre-pull tasks and confirm none exist
+REQ-2: Add documentation comment to pre_pull_images.yml stating it should not be used for CI runner container images (already cached by runner setup) or images molecule tests pull themselves
+REQ-3: Confirm gitea_runner_pre_pull_images default remains empty ([]) which is correct
+
+## Test Plan
+- Grep all molecule prepare.yml files for pre-pull patterns confirms zero matches
+- Verify pre_pull_images.yml comment is present and accurate
+- Verify gitea_runner_pre_pull_images default is [] in defaults/main.yml
+- Run make lint-ci to confirm no lint regressions
+
+## Deploy Plan
+- Merge to master via auto-merge workflow
+- No runtime changes; documentation-only
+
+## Rollback Plan
+- Revert the merge commit; comments are removed, no functional impact
+
+## Acceptance Criteria
+- [x] REQ-1: No molecule prepare.yml files in the grm repo contain pre-pull tasks (audit confirmed via grep)
+- [x] REQ-2: pre_pull_images.yml contains a comment documenting it should not be used for CI runner container images or images molecule tests pull themselves
+- [x] REQ-3: gitea_runner_pre_pull_images default remains empty ([]) in defaults/main.yml

docs/specs/GRM-173.md

@@ -0,0 +1,46 @@
+# GRM-173: Add dependency-graph, deployment-coordination, and skill-creation skills
+
+## Problem
+Agents working across the oblachno ecosystem lack shared, written context
+for three recurring struggles: (1) knowing which repo produces what and
+the correct order for cross-repo changes, (2) coordinating grm releases
+with the downstream infra dependency PR, and (3) creating and validating
+new Devin skills consistently. Without these skills, agents repeatedly
+make mistakes such as deploying infra before the grm dependency PR is
+merged, or writing skills that fail the validator.
+
+## Approach
+Add three skill files under `.devin/skills/`. Two are shared skills
+(`dependency-graph`, `skill-creation`) that must be identical across
+repos; one is grm-specific (`deployment-coordination`). All three
+follow the standard skill structure (H1 title, When to Invoke,
+Prerequisites, core content) and reference real make targets, file
+paths, and API endpoints.
+
+REQ-1: Add `.devin/skills/dependency-graph/SKILL.md` — shared skill mapping the oblachno ecosystem (repos, produces/consumers, dependency chain, correct change order, state verification)
+REQ-2: Add `.devin/skills/deployment-coordination/SKILL.md` — grm-specific skill covering release flow, downstream consumer, coordinating a grm change, and common mistakes
+REQ-3: Add `.devin/skills/skill-creation/SKILL.md` — shared skill for creating, validating, and maintaining skills (structure, quality standards, scope rules, automated validation, checklist)
+
+## Files Affected
+- `.devin/skills/dependency-graph/SKILL.md` (new)
+- `.devin/skills/deployment-coordination/SKILL.md` (new)
+- `.devin/skills/skill-creation/SKILL.md` (new)
+- `docs/specs/GRM-173.md` (new)
+
+## Test Plan
+- Verify all three SKILL.md files follow the required structure (H1, When to Invoke, Prerequisites)
+- Verify referenced make targets and file paths are accurate
+- Run `make pytest-cov` to confirm no test regressions (skills are docs-only, no code changes)
+- Confirm shared skills (`dependency-graph`, `skill-creation`) are ready for cross-repo sync
+
+## Deploy Plan
+- Merge to master via auto-merge workflow
+- No runtime changes; documentation-only (`.devin/**` is infrastructure path, no release triggered)
+
+## Rollback Plan
+- Revert the merge commit; skill files are removed, no functional impact
+
+## Acceptance Criteria
+- [x] REQ-1: `.devin/skills/dependency-graph/SKILL.md` exists with ecosystem map, dependency chain, correct change order, and state verification sections
+- [x] REQ-2: `.devin/skills/deployment-coordination/SKILL.md` exists with release flow, downstream consumer table, coordination steps, and common mistakes
+- [x] REQ-3: `.devin/skills/skill-creation/SKILL.md` exists with skill structure template, quality standards, scope rules, automated validation, and creation checklist

src/grm/__init__.py

@@ -1,3 +1,3 @@
 """Gitea Runner Manager — lean CLI for managing Gitea Actions runners."""
 
-__version__ = "0.22.1"
+__version__ = "0.23.0"