Keep the source-bound fixture regression, runner preservation proof, required map synchronization and pending full gates in the existing dated record.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Reject restrictive or unresolved ACP network settings with field-specific Legacy guidance instead of silently discarding them. Preserve unused CLI flags, recorded native profiles and trusted server policy. Synchronize the required feature and access maps without claiming live qualification.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Copy the immutable target JSON into the existing isolated helper fixture without changing negative assertions. Synchronize environment qualification recipes and access limits under the updated map contract.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Retain source-bound regression and preview evidence, current pending gates, and the independent prerequisite split.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Apply Legacy ACP alias precedence before native Codex compatibility checks. Project actual import edits onto the saved native configuration, preserving hidden runtime and credential settings while honoring explicit removals and harness changes.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Keep selected and saved effort in the existing native builder while excluding Codex effort from other providers. Update the existing builder contract and regression controls.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Explicit null clears user metadata. Preserve only the current database proof during metadata and status saves, and invalidate it atomically on config or driver updates. Real PostgreSQL interleaving controls protect concurrent qualification writes.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Keep the discovery fixture truthful for unknown platforms and add a measured GNU2.39 ARM64 control. Record existing managed fixture, cost reporting and route qualification evidence without claiming missing live acceptance.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Use the existing environment probe before native-generation Codex creation and inspect automatic server defaults. Include only the owned setup probe lease in the existing runtime estimator. Preserve explicit and Legacy cases, cleanup, unknown failure costs and reporting.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Keep explicit Legacy permission choices through automatic selection, import editing, draft reset and policy display. Use company-scoped measured immutable-image facts for automatic sandbox defaults without allocating resources during creation or discovery.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Reject named template bindings over an immutable base image. Preserve the measured company and custom-image identity without claiming compatibility for a different boot source.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Co-Authored-By: Paperclip <noreply@paperclip.ing>
* codex/runner-native-platform-prerequisites:
fix(runner): close review gaps in setup and platform qualification
Read inert exact-source target data before selecting hosted native jobs. Bound registration teardown and retain unconfirmed recovery state. Record company-scoped sandbox platform facts only for immutable OCI images after confirmed cleanup; preserve existing connection and runner behavior.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
SSH already persists suspended harness backups. Permit the same integrity and full-scope verification used by sandbox targets during a cold controller continuation.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
SSH already persists suspended harness backups. Permit the same integrity and full-scope verification used by sandbox targets during a cold controller continuation.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Extract the tested runtime and installed-package prerequisite without changing agent creation defaults or experimental gates.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Contributors use the feature map to understand and verify product
capabilities.
> - Agents need APIs and usable entry points within their permission
boundaries.
> - Product recipes and agent access need separate records with shared
capability IDs.
> - This pull request establishes that contract and gives maintenance
ownership of both maps.
> - The benefit is repeatable access review without changing product
behavior or permissions.
## Linked Issues or Issue Description
Refs: #15983, which added the portable verification skills and updated
the feature map. This is its contract-only follow-up. Related: #15288
added the original feature map.
**Issue type**
Missing documentation and maintenance contract.
**Where is the issue?**
The product map, its missing agent access companions, and the
verification skills.
**What's wrong?**
There is no shared convention for API presence, agent entry points,
permission boundaries, access gaps, or evidence. An API declaration does
not prove that a permitted native or remote agent can use it.
**Suggested fix**
Keep the product and agent access maps as siblings under `maps/`. Reuse
capability IDs. Record representative source inventory separately from
operation review and live proof. Keep both maps aligned through the
maintenance skill.
## What Changed
- Move the existing product map to `maps/feature-map/`. Preserve its
recipes and Poteto format. Update relative links and add companion
links.
- Add `maps/agent-access-map/` with 35 area files and 171 matching
capability IDs. Seed representative API operations, agent paths,
policies and known gaps from the reviewed source inventory. Mark entries
as `baseline`, `source-only`, and live verification `not-run`.
- Define the shared contract in `maps/README.md`. Track API presence,
agent availability, policy, evidence and review depth separately.
Require operation enumeration before marking a broad capability
operation-reviewed.
- Extend `maintain-verification-skill` to reconcile both maps and
inspect actual route policy and production tool admission. Update
creation, verification, notices and contributor references to use the
shared contract.
- Add `scripts/check-maps.mjs` and ten regression tests. Check matching
areas/IDs, indexes, references, contract fields, API method/path
examples, justified exceptions and evidence-state consistency.
- Keep the coverage inventory unchanged. Preserve skill invocation
policies and both MIT license blocks. No product API, permission, UI,
runtime helper, migration, dependency or workflow changes.
## Verification
- Passed: `node scripts/check-maps.mjs` reports 35 paired areas and 171
matching capability IDs. All local map references resolve.
- Passed: `node --test scripts/check-maps.test.mjs` runs ten tests.
Fixtures cover missing companions, renamed/orphan/duplicate IDs, broken
references, missing operation fields, empty exceptions, missing API
examples, empty policy fields, and verified evidence without passed
verification.
- Passed: metadata validation for all three verification skills. Skill
and contributor links resolve. Invocation policies and both license
blocks are unchanged.
- Compared relocated recipes with current master. Contents differ only
in relative links and companion links. `coverage.json` is byte-for-byte
unchanged. Master's updated read-only interaction test reference is
retained.
- Passed: `git diff --check`. The branch starts from current master and
contains only this contract change.
- Repository-wide typecheck, tests and build were not run locally for
this documentation and Node structural-check scope. Live agent access
was not exercised. Those gaps are explicit in the baseline entries; the
structural check does not certify behavior. CI and automated review are
pending.
## Risks
- The companions are representative source baselines. They are not
complete operation inventories or passed agent journeys.
- The checker cannot establish actual permissions, native tool
admission, operation completeness or truthful live evidence. Maintenance
must inspect source and verify with the actual permitted agent context.
- External links to the old product-map directory may need updates.
Repository references have been updated.
- Existing access gaps and board/user restrictions remain recorded. This
contract does not change authority or implement missing paths.
## Model Used
OpenAI GPT-6 through Codex assisted with this change. The session does
not expose the exact serving model ID, configured context-window size,
or reasoning setting. Capabilities used: source analysis, file edits,
shell commands, and structural regression tests.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
Use the controller override only for matching remote OS and CPU, preserving explicit remote priority and strict target header verification.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Record stop-time ACP turn diagnostics (last-event age, oldest pending tool age, pending tool kinds) so wall-clock timeout events show whether the turn was stalled. Diagnostics only.
Co-Authored-By: Paperclip <Paperclip-Paperclip@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
## Thinking Path
> - Paperclip manages AI agents and their work.
> - Chat channels connect that work to provider conversations.
> - Provider workflows still fill the shared service.
> - Their authority and recovery rules need clear homes.
> - This PR lands five reviewed extractions into channel folders.
> - Shared authorization, state and worker lifetime stay explicit.
## Linked Issues or Issue Description
Refs #15976, #15996, #15999, #16001 and #16002. This root PR now
contains all five extractions from batch three. The four upper PRs were
merged into their parents. The combined tree matches reviewed stack head
`789e918b5af49ec1958076ce4d22d31006205302` exactly.
## What Changed
- Extract Slack session status and accepted-file receipt recovery.
- Extract Telegram publication drafts and Stop handling.
- Extract Photon events and native responses.
- Extract Teams file-transfer authority.
- Reduce `chat-channels.ts` from 29,897 to 26,461 lines (3,436 lines,
11.5%).
- Preserve function bodies, runtime-map identity, leases and transaction
boundaries.
## Verification
- All 117 selected public-service cases passed before and after their
respective extractions.
- All 30 moved declarations and remaining service statements preserve
normalized syntax across the five slices.
- Server typecheck, module boundaries and whitespace checks passed.
- Cumulative `pnpm -r typecheck` and `pnpm build` passed. Combined-head
CI and review are pending on `5de1b0b99d71049a72440e8c131bf4bb30bdf3f3`.
- Fixtures exercise assembled services and disposable databases, not
live third-party providers.
- Cumulative local `pnpm test:run` on
`789e918b5af49ec1958076ce4d22d31006205302` exited 1: 32,414 passed, 125
skipped, one failure. `company-skills-routes.test.ts:1691` (`stars,
forks, and comments on skills through company-scoped endpoints`) failed
with `socket hang up`; that test and its route are unchanged by this
stack. The same 58-case suite then passed in isolation without code
changes. The original command did not reach 111 later serialized suites;
its failure is retained, so the full-local-test checklist remains
unchecked.
```sh
pnpm exec vitest run server/src/__tests__/chat-channels.integration.test.ts -t 'Slack session|Slack turns to Idle|signed Slack session stop|denies guest Slack stops'
```
- Combined-head CI attempt 1: five runners received a shutdown signal at
19:13:08 UTC (server shard 10/12, serialized shards 2/9 and 7/9, E2E
shards 3/8 and 7/8). These jobs stopped during execution without a
failing test assertion. [Original
workflow](https://github.com/paperclipai/paperclip/actions/runs/38293566028).
The affected jobs will receive one unchanged retry after the other jobs
finish.
## Risks
- Changes cross durable provider recovery and authority boundaries. The
moved bodies preserve syntax, and 117 focused public-service cases
passed before and after their extractions.
- Shared workers, timers, runtime maps and generic authorization remain
in the service.
- The full local test run failed in an unchanged company-skills route
test; its result and the unrun suites are recorded above.
## Model Used
OpenAI Codex, GPT-6 family, with reasoning and repository tools. The
exact runtime model ID and context-window size are not exposed in this
session. No separate agent was used.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [ ] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
Retain both hosted cache scopes. Use shared test-home teardown and restore owned directory permissions for the observed Mac ENOTEMPTY failure while keeping unrelated errors visible.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Contributors use the feature map to select verification recipes.
> - The repository needs separate skills to create a verifier, maintain
recipes, and run checks.
> - Shared recipes must match current routes, controls, and
prerequisites.
> - This pull request adds portable verification skills and updates the
existing map.
> - The benefit is a repeatable verification workflow with clear
ownership and evidence limits.
## Linked Issues or Issue Description
Refs: #15288, which added the existing feature map. The related test
cleanup in #15978 changes test ownership; this PR changes skill
instructions and recipes.
**Issue type**
Missing documentation and outdated verification instructions.
**Where is the issue?**
The project-local skills in `.agents/skills/` and the recipes in
`feature-map/`.
**What's wrong?**
The repository lacks the portable creation and maintenance skills. The
verifier mixes live verification with map editing. Some recipes describe
old controls or omit required flags, identities, and provider setup.
**Suggested fix**
Import the skills and their complete references from Paperclip's pstack
fork. Keep the existing map in its current directory. Let the verifier
read recipes and report drift. Let the maintenance skill update the map.
## What Changed
- Add `create-verification-skill` and `maintain-verification-skill` from
the portable fork of Poteto's pstack. Include invocation policies, all
feature-map examples, self-contained attribution notices and the full
MIT license in each skill folder.
- Use the existing `feature-map/` directory and Poteto's recipe format.
Omit the fork's separate runtime skill and runtime-skill dependencies
because Paperclip supplies execution.
- Add `verify-paperclip` with executable launch, readiness, onboarding,
and cleanup helpers. Use isolated temporary data and separate evidence
directories.
- Make `verify-paperclip` consume the map and report recipe drift. Give
map editing to `maintain-verification-skill`.
- Review all 35 feature files and correct 30 recipes. Clarify routes,
flags, current controls, persistence scope, and the limits of linked
tests.
- Remove the unrelated Conference Room flag change from the onboarding
smoke helper.
- Add a standard-library ownership test. It proves that doctor and stop
reject a foreign checkout without stopping its live server. A
delayed-exit fixture also proves cleanup waits for runtime exit.
- Keep audit screenshots and scratch reports outside the repository. No
product code, migrations, lockfile, or workflow changes.
## Verification
- Passed after rebase: metadata validation for all three skills, Python
syntax and executable checks, JavaScript syntax, 35-entry map/index
consistency, relative links, license preservation, and `git diff
--check`.
- Ran one isolated live audit before rebase. Checked task persistence,
document upload/download bytes, goal hierarchy, project edits, search,
local process Agent Chat, question replies, queued-message discard/stop,
and a harmless loopback MCP call.
- Checked local settings and draft flows. Synthetic cost events reached
a small budget limit and paused the test agent. These checks do not
prove external model quality or provider billing.
- Checked cleanup of the audit server, database, local MCP fixture, and
workspace service inventory.
- Passed: `pnpm -r typecheck`.
- Greptile: 5/5 on commit `4344ac4df7`. The cleanup finding is fixed and
its review thread is resolved. A fresh review is pending for the notice
relocation.
- Passed: `python3 -m unittest discover -s
.agents/skills/verify-paperclip/tests -v`. Both tests pass: the foreign
fixture stayed alive after rejection, and owned cleanup waited for
delayed runtime exit.
- Passed: `pnpm build`.
- All GitHub CI checks were green on commit `4344ac4df7`. The cloud-auth
browser shard passed after one retry for a disposed mocked API response.
The notice relocation passed skill metadata, relative-link, and
license-preservation checks; fresh CI/review is pending.
- The full local `pnpm test:run` did not complete. It was interrupted
after nearly 30 minutes once the complete CI suite passed on the same
commit. Exit code: 130. The new helper safety tests passed locally. The
local full-suite checklist item stays unchecked.
- Full live coverage is incomplete. Authenticated multi-user checks and
external provider journeys need their own identities, credentials, and
targets.
- The corrected onboarding helper could not be re-tested after removing
the flag mutation. Two fresh launches failed before server startup.
Direct PostgreSQL initialization reported exhausted shared-memory IDs or
limits. The host had free disk space. No host-wide limits or unrelated
instances were changed.
## Risks
- Recipe drift can return as the UI changes. The map records
verification scope, not proof of a passed journey.
- The new helpers assume a source checkout with installed dependencies,
Python, Node, and `lsof`. The onboarding helper also needs Playwright.
- The local audit found observations that need separate investigation:
an action test returned fixture data while the connection control showed
Off and native profiles existed; a failed status-card setup task left
the card showing Setting up. Workspace service start also did not reach
a running state. This PR does not change product behavior or redefine
those outcomes as success.
- The main audit trace did not finalize during browser shutdown.
Screenshots, text captures, and iteration results were retained.
- Live evidence was captured before rebase. The onboarding helper
re-test and full provider/auth qualification remain incomplete.
## Model Used
OpenAI GPT-6 through Codex assisted with this change. The session does
not expose the exact serving model ID, configured context-window size,
or reasoning setting. Capabilities used: source analysis, read-only
source subagents, file edits, shell commands, and browser automation.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [ ] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Its trusted PR workflow builds the Runner's Rust binary in five
jobs: Typecheck + Release Registry, Build, Canary Dry Run, and the
`rust` and `vitest 2/2` lanes of Verify Paperclip Runner. The `Build`
job is the slowest check in almost every green run (461s, 489s and 515s
in runs 38180574768, 38182717052 and 38183149920 on 2026-10-11).
> - Each of those jobs restores master's `release-runner-v2` dependency
cache. That cache holds only third-party crates. `rust-cache` removes
workspace crates when it saves, and a fresh checkout invalidates cargo's
mtime fingerprints for workspace crates in any case. So cargo compiles
`paperclip-runner-core` in release mode from scratch on every run. The
job logs show one `Compiling` line and 28s to 49s for it.
> - 18 of the 239 commits on `master` in the seven days before this
change touched the crate's inputs. The other 221 compiled the same
sources to the same binary again.
> - This pull request adds a second cache that holds only that crate's
release artifacts (about 70 MB). The key is made from the compiler, the
checkout root and a hash of every input the build reads. The master Rust
lane writes it. A pull request whose merge tree has byte-identical
inputs restores it and marks the artifacts newer than the checkout, so
cargo keeps them. Every other pull request misses and compiles as
before.
> - The benefit is about 28s to 49s less on `Build`, which is the run's
critical path, and the same on Typecheck, Canary Dry Run and two Runner
lanes, for about nine of ten pull requests.
## Linked Issues or Issue Description
No GitHub issue exists for this. The description below follows the
enhancement template.
**What existing behavior does this improve?**
The time the trusted PR workflow spends to compile the Runner's own
crate on every run.
**Subsystem affected**
CI / build
**Current behavior**
Every PR job that builds the Runner binary compiles
`paperclip-runner-core` in release mode from scratch. The compile takes
28s to 49s on the `runs-on` fleet (one `Compiling paperclip-runner-core`
line, then `Finished release profile` in runs 38169436532 to 38183149920
on 2026-10-11). The shared Rust cache cannot help: it holds only
third-party crates, and a fresh checkout gives every source a new mtime,
which makes cargo discard the crate's fingerprints.
**Proposed behavior**
The master Rust lane in `release-verify.yml` saves the crate's release
artifacts under a key made from the compiler, the checkout root and a
hash of every input the release build reads. The PR jobs that build the
binary restore that entry read-only. On a hit, a script sets the mtime
of every restored artifact to now, so cargo sees every source as older
than the build and keeps the artifacts. `pnpm build` then prints
`Finished release profile` in under a second for that crate. On a miss,
cargo compiles the crate exactly as before.
**Reason and benefit**
About 28s to 49s off `Build`, the slowest check in almost every green
run, and the same off the Typecheck, Canary Dry Run and two Verify
Paperclip Runner lanes. The saving applies to every pull request whose
merge tree does not change the crate's inputs, which was 221 of 239
`master` commits in the week before this change.
**Breaking changes**
None. The checks are unchanged. A miss or an eviction costs the compile
time it costs today and nothing else.
**Additional context**
This is a workflow change. `pr.yml` calls `pr-trusted.yml@master`, so
this pull request's own run uses the current reader steps, and the
writer runs only on a master push. The saving becomes visible on the
first PR runs after the merge and after the next master push has written
the entry.
## What Changed
- `scripts/runner-release-crate-cache.mjs` (new): the one place that
defines the cached path list (`CACHED_PATHS`), the hashed inputs
(`HASHED_INPUTS`) and the files a restore must contain (`REQUIRED`). The
`scope` subcommand prints the platform, the `rustc --version --verbose`
release, commit and host, and a hash of the checkout root. The
`mark-current` subcommand sets the mtime of every restored artifact to
now and prints a `paperclip.runner-release-crate.v1` JSON line with the
file count and bytes. If the restored set lacks any of the seven files
cargo's freshness check reads or checks for (the fingerprint hash and
dep-info of each unit, the lib's rlib and rmeta, the binary under
`deps`), it removes the restored files, prints `action: discarded`
naming the missing file, and exits 0, so cargo compiles as on a miss.
- `.github/workflows/pr-trusted.yml`: after the `release-runner-v2`
restore in `typecheck_release_registry`, `build` and `canary_dry_run`,
and in the `rust` and `vitest 2/2` lanes of `verify_paperclip_runner`
(new matrix key `release_crate: true`), three steps: compute the key
scope, `actions/cache/restore` the crate under
`runner-release-crate-v1-<scope>-<hashFiles(...)>` with no
`restore-keys`, and run `mark-current` when `cache-hit == 'true'`. The
`hashFiles` list covers the toolchain pin, the manifests and lockfile,
the crate sources, the protocol tree and the three fixture trees the
test modules include (two under the Runner package and
`tests/runner-e2e/fixtures` at the repository root).
- `.github/workflows/release-verify.yml`: in the Rust lane of an
own-repository master push, after the debug warm-up: compute the key
scope, look the key up (`lookup-only: true`), and on a miss run `cargo
clean --release -p paperclip-runner-core` and `pnpm build:binary` (the
exact command the readers run, so the entry carries no test-profile
variants), then `actions/cache/save` the same path list under the same
key.
- `scripts/__tests__/runner-release-crate-cache.test.mjs` (new, 10
tests, in `test:release-registry`): the path list and hashed inputs are
the expected ones and exist; every `include_str!` and `include_bytes!`
in the crate sources resolves under a hashed input, whether its argument
is a literal or a `concat!` of literals, `env!("CARGO_MANIFEST_DIR")`
and macro metavariables, and any other form fails the test by name (this
guard found the `test/fixtures` tree during development and the
`tests/runner-e2e/fixtures` tree in review); the parser itself is tested
on a synthetic crate for each supported form and six rejected ones;
every PR reader restores after the dependency cache and marks before its
build command with the script's exact path list and key and no
`restore-keys`; the Verify lanes are gated by `release_crate`; the
writer runs after the checks with the lookup, the clean and build, and
the save under the readers' exact path list and key; the scope changes
with the compiler and the checkout root and survives a missing `rustc`;
`mark-current` makes every artifact newer than the checkout, leaves a
dependency's artifacts alone, discards a restore missing any one of the
seven required files (one case each, checking the named reason), and
still marks a restore missing any one of the seven files cargo does not
need.
## Review fixes
Greptile's two findings on the first commit were both valid and are
fixed in the second commit:
- **Incomplete caches looked usable.** `REQUIRED` checked four files,
but cargo's freshness check needs seven. Measured with cargo 1.97.1 on
the real crate (restore, touch the artifacts, delete one file, build):
without the lib's fingerprint hash, dep-info, rlib or rmeta it
recompiles the crate (24s to 27s); without the bin's fingerprint hash,
dep-info or the binary under `deps` it relinks (1.5s); without either
`.json` fingerprint record, either rustc `.d` file, the uplifted binary
or its `.d` it reports the crate fresh in 0.15s and recreates the
uplifted binary from `deps`, and the staging step then copies it.
`REQUIRED` now lists exactly the seven files, the glob for the binary
under `deps` excludes the `.d` beside it, and the test removes each of
the seven in turn (discarded, naming it) and each of the other seven in
turn (still marked).
- **Composed include paths escaped the guard.** The regex matched only
`include_str!("literal")` and skipped `include_str!(concat!(...))`,
which the crate uses four times (30 of 34 includes were checked). The
guard now parses every include argument and resolves the `concat!`
forms; one of them, in `provider_events.rs`, reads
`tests/runner-e2e/fixtures/native-completion/terminal-tool-carrier.json`
at the repository root, outside every hashed input. It is test-only, so
today's release binary does not read it, but the key is meant to cover
every include whatever its build, so `tests/runner-e2e/fixtures/**` (9
files, 80 KB) joins `HASHED_INPUTS` and the six key embeddings in the
two workflows. With the first commit's input list, the new guard fails
on exactly that include.
- `doc/RELEASE-AUTOMATION-SETUP.md`: a "Runner release crate cache"
section under the Runner verification dependency cache, with the key,
the writer, the readers and the trust boundary.
- `package.json`: the new test file in `test:release-registry`.
## Verification
- `node --test scripts/__tests__/runner-release-crate-cache.test.mjs`:
10 tests pass, three consecutive runs.
- The second commit, with the real toolchain on the real crate (`rustc
1.97.1`, cold build 2m20s, the crate's artifacts 74 MB): every source
touched to a fresh-checkout mtime and the archived artifacts restored,
cargo compiles (`Compiling paperclip-runner-core`, 24.8s and 25.1s, the
two controls); restored and marked by `mark-current`, `Finished release
profile in 0.16s` with no `Compiling`, and the same on a second build.
Then, for each of the fourteen files cargo wrote for the two units, one
at a time: restored, marked and that file deleted, cargo recompiles
without any of the seven `REQUIRED` files and reports fresh without any
of the other seven; restored with that file deleted and run through
`mark-current`, the seven required ones print `action: discarded` naming
the file and the others print `marked`. A restore holding only the seven
required files builds fresh in 0.15s, cargo recreates the uplifted
binary, and `stage-runner-binary.mjs` copies it into `dist/bin`.
- `node --test '.github/scripts/tests/*.test.mjs'`: 378 of 381 pass. The
3 failures are in `docker-canary-promotion.test.mjs` and fail the same
way on an unmodified `master` checkout in this sandbox (the stand-in
docker writes no `calls.jsonl`).
- `node --test scripts/__tests__/release-verify-workflow.test.mjs
scripts/__tests__/e2e-shard.test.mjs
scripts/__tests__/run-vitest-stable-shard.test.mjs`: 99 of 101 pass; the
2 failures spawn `pnpm`, which is not on this sandbox's PATH, and fail
the same way on `master`.
- Both workflows parse as YAML. No job has a duplicate step name or step
id.
- The mechanism, with the real toolchain (`rustc 1.97.1`, the package's
pin) on a copy of the crate: a cold `cargo build --release -p
paperclip-runner-core --bin paperclip-runnerd` (2m23s here). Then the
artifact set was archived with `tar`, removed, every source re-copied so
each had a new mtime, and the set restored:
- restored, not marked: `Compiling paperclip-runner-core`, 24.6s (the
control; this is what CI does today);
- restored and marked by `mark-current`: `Finished release profile in
0.15s`, no `Compiling`; a second build is the same;
- one Rust source edited after a marked restore: `Compiling`, 24.0s;
- one protocol schema edited after a marked restore: `Compiling`, 23.9s
(the protocol files enter cargo's dep-info as absolute paths and are
hashed into the key);
- the rlib removed from a restored set: `mark-current` prints `action:
discarded, reason: incomplete: the lib rlib is missing` and removes the
rest, and cargo compiles.
- Not run here: the workflow steps themselves. `pr.yml` pins
`pr-trusted.yml@master`, so this pull request's own `Build` job runs the
current reader steps, and the writer runs only on a master push. The
first PR runs after the merge and after the next master push show the
`Restore the Runner release crate` step's result and the
`paperclip.runner-release-crate.v1` line in their `Build` logs.
## Risks
- Correctness: a reader can only restore artifacts that master built
from inputs byte-identical to its own merge tree, because every input
the build reads is in the key, and the key has no prefix match. A pull
request cannot write. The trust boundary is unchanged: master writes, a
pull request restores master's entry.
- An input the build reads that is outside the hashed patterns would let
a stale artifact pass. The release dep-info lists the crate sources and
the protocol schemas and fixtures; all are hashed. The guard test
resolves every `include_str!` and `include_bytes!` in the crate, literal
or `concat!`, and fails if one lands outside the hashed inputs or uses a
form it cannot resolve, so a new include form cannot slip past it by not
matching a regex.
- GitHub-hosted runners check out under `/home/runner/work`, the
`runs-on` fleets under `/home/runner/_work`. Cargo records the included
protocol files as absolute paths, so an entry from one root is useless
under the other. The checkout root is in the key, so the other root
misses instead of restoring something cargo then discards.
- A cache that does not keep mtimes on restore changes nothing: the
artifacts are then already newer than the checkout, and `mark-current`
sets them to now in any case.
- Cost: one new entry of about 70 MB per distinct input set, written
only by the master Rust lane on a miss, plus one `cargo clean` and
`build:binary` (about 40s to 50s) on that lane when the inputs changed.
The readers pay one restore of about 70 MB (two to three seconds at the
fleet's observed restore speed) and a sub-second `mark-current`.
- An incomplete entry (a writer that saved a partial set) is discarded
by `mark-current` and costs one compile, the same as a miss.
## Model Used
- Claude (Anthropic), model id `claude-fable-5-1` (Claude Fable 5.1),
extended thinking, with tool use (shell, file editing, GitHub CLI, a
local Rust toolchain for the experiment) in a Paperclip agent session.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: devinfoley <139239+devinfoley@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
