feat(artifacts): define artifact storage types, config, and persistence layer#2801
Merged
M3gA-Mind merged 2 commits intoMay 28, 2026
Merged
Conversation
…ce layer Add new `artifacts` domain with ArtifactMeta struct, ArtifactKind/ArtifactStatus enums, filesystem-backed persistence under <workspace>/artifacts/<uuid>/, and three RPC endpoints (ai_list_artifacts, ai_get_artifact, ai_delete_artifact). Includes 38 unit tests covering CRUD operations, pagination, path-sandboxing enforcement, and corrupt metadata handling. Closes tinyhumansai#2776
Contributor
📝 WalkthroughWalkthroughThis PR implements the artifact storage domain for OpenHuman, introducing typed artifact metadata (ArtifactMeta), filesystem-backed persistence with path validation, and three RPC endpoints (ai_list_artifacts, ai_get_artifact, ai_delete_artifact) that enable workspace artifact management with pagination, sorting, corruption resilience, and comprehensive test coverage. ChangesArtifact Storage Domain
Sequence Diagram(s)sequenceDiagram
participant Client
participant ops
participant store
participant filesystem
Client->>ops: ai_list_artifacts(offset, limit)
ops->>store: list_artifacts(workspace_dir, offset, limit)
store->>filesystem: read artifacts/ directory entries
store->>filesystem: read each meta.json
store->>store: sort by created_at descending
store-->>ops: (Vec<ArtifactMeta>, total)
ops-->>Client: JSON with artifacts, total, offset, limit
Client->>ops: ai_get_artifact(artifact_id)
ops->>store: get_artifact(workspace_dir, artifact_id)
store->>store: validate_artifact_id
store->>store: assert_within_root
store->>filesystem: read artifact/meta.json
store-->>ops: ArtifactMeta
ops->>ops: compute absolute_path
ops-->>Client: JSON with meta + absolute_path
Client->>ops: ai_delete_artifact(artifact_id)
ops->>store: delete_artifact(workspace_dir, artifact_id)
store->>store: validate and assert containment
store->>filesystem: remove artifact directory recursively
store-->>ops: success
ops-->>Client: JSON with artifact_id, deleted: true
sequenceDiagram
participant Client
participant schemas
participant ops
Client->>schemas: RPC request with function and params JSON
schemas->>schemas: load config with timeout
schemas->>schemas: parse/validate parameters
schemas->>ops: call ai_* operation
ops-->>schemas: RpcOutcome result
schemas->>schemas: convert to JSON
schemas-->>Client: JSON response
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Comment |
coderabbitai
Bot
added
feature
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 4
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.claude/memory.md:
- Line 217: Update the note to correct the behavior and point to the canonical
policy: clarify that assert_within_root does not canonicalize but only checks
prefix containment (starts_with), remove the instruction to “replicate this
pattern” for new domains, and instead direct teams to use the centralized file
I/O validation in src/openhuman/security/policy.rs (i.e., call validate_path /
validate_parent_path) while keeping validate_artifact_id's two-layer checks as
an example for artifact IDs only.
In `@src/openhuman/artifacts/ops.rs`:
- Around line 63-69: The current derivation of absolute_path uses meta.path
directly which allows absolute paths or parent-traversal to escape the artifacts
root; update the logic that builds absolute_path (the code touching
absolute_path, config.workspace_dir, and meta.path) to first validate/sanitize
meta.path: reject or canonicalize if Path::new(&meta.path).is_absolute() and
reject or disallow any Path::new(&meta.path).components() containing
Component::ParentDir, or alternatively rebuild a safe relative PathBuf by
filtering components to only Normal components before joining; if validation
fails return an error instead of joining, then join the sanitized relative path
to config.workspace_dir.join("artifacts") and convert to string as before.
In `@src/openhuman/artifacts/schemas.rs`:
- Around line 87-92: The output schema for get_artifact currently uses
TypeSchema::Ref("ArtifactMeta") but the comment claims an extra absolute_path
field; update the contract to match by replacing the output FieldSchema type
with an explicit schema that includes ArtifactMeta plus absolute_path (either by
changing the type to Ref("ArtifactMetaWithPath") and adding that new type
definition, or by using an inline TypeSchema::Object that spreads ArtifactMeta
fields and adds absolute_path: TypeSchema::String required: true); ensure the
FieldSchema name stays "artifact" and adjust any schema registry entries that
reference ArtifactMeta if you introduce ArtifactMetaWithPath.
In `@src/openhuman/artifacts/store.rs`:
- Around line 26-57: The validator validate_artifact_id currently allows the
single-dot "." which leads to delete_artifact calling remove_dir_all on the
artifacts root; update validate_artifact_id to reject "." (and also reject "./"
and ".\\" to be safe) by returning an Err with a clear message (same format as
other errors), so delete_artifact and any other callers cannot use "." as an
artifact_id and accidentally operate on the whole artifacts tree.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 37ca5f18-69fc-4ea8-ad37-543ac0921a5a
📒 Files selected for processing (10)
.claude/memory.mdsrc/core/all.rssrc/openhuman/artifacts/mod.rssrc/openhuman/artifacts/ops.rssrc/openhuman/artifacts/ops_tests.rssrc/openhuman/artifacts/schemas.rssrc/openhuman/artifacts/store.rssrc/openhuman/artifacts/store_tests.rssrc/openhuman/artifacts/types.rssrc/openhuman/mod.rs
|
|
||
| - **Filesystem-backed persistence, no SQLite** — `src/openhuman/artifacts/` stores JSON metadata (`meta.json`) + binary blobs under `<workspace_dir>/artifacts/<uuid>/`. Pattern mirrors `memory/ops/files.rs` but simpler. | ||
| - **`"ai"` namespace in controller registry** — RPC methods are `openhuman.ai_list_artifacts`, `openhuman.ai_get_artifact`, `openhuman.ai_delete_artifact`. Future `ai_*` methods should use this same namespace. | ||
| - **Two-layer path validation required** — (1) `validate_artifact_id` rejects empty strings, `/`, `\`, `..`, absolute Unix paths, Windows `C:` and UNC `\\` paths; (2) `assert_within_root` canonicalizes and checks containment. Replicate this pattern for any new filesystem-backed domain. |
Contributor
There was a problem hiding this comment.
Correct the security guidance and assert_within_root description.
This line is currently misleading in two ways: assert_within_root here does not canonicalize paths (it checks starts_with), and advising teams to “replicate this pattern” conflicts with the documented policy that file I/O validation must go through src/openhuman/security/policy.rs (validate_path / validate_parent_path). Please update this note to avoid steering future domains toward a non-policy pattern.
Suggested doc fix
-- **Two-layer path validation required** — (1) `validate_artifact_id` rejects empty strings, `/`, `\`, `..`, absolute Unix paths, Windows `C:` and UNC `\\` paths; (2) `assert_within_root` canonicalizes and checks containment. Replicate this pattern for any new filesystem-backed domain.
+- **Artifacts path safety (domain-specific)** — this domain currently uses (1) `validate_artifact_id` input checks and (2) `assert_within_root` containment checks before filesystem operations. Note: `assert_within_root` is a containment check (not canonicalization). For new filesystem-backed domains, follow the security policy entry points in `src/openhuman/security/policy.rs` (`validate_path` / `validate_parent_path`).📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| - **Two-layer path validation required** — (1) `validate_artifact_id` rejects empty strings, `/`, `\`, `..`, absolute Unix paths, Windows `C:` and UNC `\\` paths; (2) `assert_within_root` canonicalizes and checks containment. Replicate this pattern for any new filesystem-backed domain. | |
| - **Artifacts path safety (domain-specific)** — this domain currently uses (1) `validate_artifact_id` input checks and (2) `assert_within_root` containment checks before filesystem operations. Note: `assert_within_root` is a containment check (not canonicalization). For new filesystem-backed domains, follow the security policy entry points in `src/openhuman/security/policy.rs` (`validate_path` / `validate_parent_path`). |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In @.claude/memory.md at line 217, Update the note to correct the behavior and
point to the canonical policy: clarify that assert_within_root does not
canonicalize but only checks prefix containment (starts_with), remove the
instruction to “replicate this pattern” for new domains, and instead direct
teams to use the centralized file I/O validation in
src/openhuman/security/policy.rs (i.e., call validate_path /
validate_parent_path) while keeping validate_artifact_id's two-layer checks as
an example for artifact IDs only.
Comment on lines
+63
to
+69
| // Compute absolute path for the caller's convenience. | ||
| let absolute_path = config | ||
| .workspace_dir | ||
| .join("artifacts") | ||
| .join(&meta.path) | ||
| .to_string_lossy() | ||
| .into_owned(); |
Contributor
There was a problem hiding this comment.
Harden absolute_path derivation against unsafe meta.path.
meta.path is used directly in join(...). If it is absolute (or contains parent traversal), the computed absolute_path can escape the artifacts root.
Suggested fix
+ use std::path::{Component, Path};
+
+ let rel = Path::new(&meta.path);
+ if rel.is_absolute()
+ || rel
+ .components()
+ .any(|c| matches!(c, Component::ParentDir | Component::RootDir | Component::Prefix(_)))
+ {
+ return Err(format!(
+ "[artifacts] invalid artifact path for id={artifact_id}: {:?}",
+ meta.path
+ ));
+ }
+
+ let artifacts_root = config.workspace_dir.join("artifacts");
let absolute_path = config
- .workspace_dir
- .join("artifacts")
+ .workspace_dir
+ .join("artifacts")
.join(&meta.path)
.to_string_lossy()
.into_owned();📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| // Compute absolute path for the caller's convenience. | |
| let absolute_path = config | |
| .workspace_dir | |
| .join("artifacts") | |
| .join(&meta.path) | |
| .to_string_lossy() | |
| .into_owned(); | |
| // Compute absolute path for the caller's convenience. | |
| use std::path::{Component, Path}; | |
| let rel = Path::new(&meta.path); | |
| if rel.is_absolute() | |
| || rel | |
| .components() | |
| .any(|c| matches!(c, Component::ParentDir | Component::RootDir | Component::Prefix(_))) | |
| { | |
| return Err(format!( | |
| "[artifacts] invalid artifact path for id={artifact_id}: {:?}", | |
| meta.path | |
| )); | |
| } | |
| let artifacts_root = config.workspace_dir.join("artifacts"); | |
| let absolute_path = config | |
| .workspace_dir | |
| .join("artifacts") | |
| .join(&meta.path) | |
| .to_string_lossy() | |
| .into_owned(); |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@src/openhuman/artifacts/ops.rs` around lines 63 - 69, The current derivation
of absolute_path uses meta.path directly which allows absolute paths or
parent-traversal to escape the artifacts root; update the logic that builds
absolute_path (the code touching absolute_path, config.workspace_dir, and
meta.path) to first validate/sanitize meta.path: reject or canonicalize if
Path::new(&meta.path).is_absolute() and reject or disallow any
Path::new(&meta.path).components() containing Component::ParentDir, or
alternatively rebuild a safe relative PathBuf by filtering components to only
Normal components before joining; if validation fails return an error instead of
joining, then join the sanitized relative path to
config.workspace_dir.join("artifacts") and convert to string as before.
Comment on lines
+87
to
+92
| outputs: vec![FieldSchema { | ||
| name: "artifact", | ||
| ty: TypeSchema::Ref("ArtifactMeta"), | ||
| comment: "Artifact metadata plus absolute_path field.", | ||
| required: true, | ||
| }], |
Contributor
There was a problem hiding this comment.
get_artifact output schema does not encode the documented absolute_path field.
The schema advertises artifact as ArtifactMeta, but the comment says it includes an extra absolute_path. That makes discovery/docs contract ambiguous for clients. Please model the extra field explicitly (or remove that claim if not returned).
Suggested schema shape
- outputs: vec![FieldSchema {
- name: "artifact",
- ty: TypeSchema::Ref("ArtifactMeta"),
- comment: "Artifact metadata plus absolute_path field.",
- required: true,
- }],
+ outputs: vec![FieldSchema {
+ name: "artifact",
+ ty: TypeSchema::Object {
+ fields: vec![
+ FieldSchema {
+ name: "meta",
+ ty: TypeSchema::Ref("ArtifactMeta"),
+ comment: "Artifact metadata.",
+ required: true,
+ },
+ FieldSchema {
+ name: "absolute_path",
+ ty: TypeSchema::String,
+ comment: "Resolved absolute file path for the artifact.",
+ required: true,
+ },
+ ],
+ },
+ comment: "Artifact metadata plus resolved absolute path.",
+ required: true,
+ }],📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| outputs: vec![FieldSchema { | |
| name: "artifact", | |
| ty: TypeSchema::Ref("ArtifactMeta"), | |
| comment: "Artifact metadata plus absolute_path field.", | |
| required: true, | |
| }], | |
| outputs: vec![FieldSchema { | |
| name: "artifact", | |
| ty: TypeSchema::Object { | |
| fields: vec![ | |
| FieldSchema { | |
| name: "meta", | |
| ty: TypeSchema::Ref("ArtifactMeta"), | |
| comment: "Artifact metadata.", | |
| required: true, | |
| }, | |
| FieldSchema { | |
| name: "absolute_path", | |
| ty: TypeSchema::String, | |
| comment: "Resolved absolute file path for the artifact.", | |
| required: true, | |
| }, | |
| ], | |
| }, | |
| comment: "Artifact metadata plus resolved absolute path.", | |
| required: true, | |
| }], |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@src/openhuman/artifacts/schemas.rs` around lines 87 - 92, The output schema
for get_artifact currently uses TypeSchema::Ref("ArtifactMeta") but the comment
claims an extra absolute_path field; update the contract to match by replacing
the output FieldSchema type with an explicit schema that includes ArtifactMeta
plus absolute_path (either by changing the type to Ref("ArtifactMetaWithPath")
and adding that new type definition, or by using an inline TypeSchema::Object
that spreads ArtifactMeta fields and adds absolute_path: TypeSchema::String
required: true); ensure the FieldSchema name stays "artifact" and adjust any
schema registry entries that reference ArtifactMeta if you introduce
ArtifactMetaWithPath.
Comment on lines
+26
to
+57
| fn validate_artifact_id(id: &str) -> Result<(), String> { | ||
| if id.is_empty() { | ||
| return Err("[artifacts] artifact_id must not be empty".to_string()); | ||
| } | ||
| if id.contains('/') { | ||
| return Err(format!( | ||
| "[artifacts] artifact_id must not contain '/': {id:?}" | ||
| )); | ||
| } | ||
| if id.contains('\\') { | ||
| return Err(format!( | ||
| "[artifacts] artifact_id must not contain '\\': {id:?}" | ||
| )); | ||
| } | ||
| if id == ".." || id.starts_with("../") || id.starts_with("..\\") { | ||
| return Err(format!( | ||
| "[artifacts] artifact_id must not be a path traversal: {id:?}" | ||
| )); | ||
| } | ||
| // Reject absolute paths (Unix /foo or Windows C:\foo / \\server\share) | ||
| if id.starts_with('/') || id.starts_with('\\') { | ||
| return Err(format!( | ||
| "[artifacts] artifact_id must not be an absolute path: {id:?}" | ||
| )); | ||
| } | ||
| // Reject Windows drive-letter paths like C: | ||
| if id.len() >= 2 && id.as_bytes()[1] == b':' { | ||
| return Err(format!( | ||
| "[artifacts] artifact_id must not be an absolute path: {id:?}" | ||
| )); | ||
| } | ||
| Ok(()) |
Contributor
There was a problem hiding this comment.
Reject . as an artifact ID to prevent full-root deletion.
validate_artifact_id allows ".". That value flows into delete_artifact, where remove_dir_all(<workspace>/artifacts/.) can delete the entire artifacts tree.
Suggested fix
fn validate_artifact_id(id: &str) -> Result<(), String> {
if id.is_empty() {
return Err("[artifacts] artifact_id must not be empty".to_string());
}
+ if id == "." || id == ".." {
+ return Err(format!(
+ "[artifacts] artifact_id must not be a reserved path component: {id:?}"
+ ));
+ }
if id.contains('/') {
return Err(format!(
"[artifacts] artifact_id must not contain '/': {id:?}"
));
}
@@
- if id == ".." || id.starts_with("../") || id.starts_with("..\\") {
+ if id.starts_with("../") || id.starts_with("..\\") {
return Err(format!(
"[artifacts] artifact_id must not be a path traversal: {id:?}"
));
}🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@src/openhuman/artifacts/store.rs` around lines 26 - 57, The validator
validate_artifact_id currently allows the single-dot "." which leads to
delete_artifact calling remove_dir_all on the artifacts root; update
validate_artifact_id to reject "." (and also reject "./" and ".\\" to be safe)
by returning an Err with a clear message (same format as other errors), so
delete_artifact and any other callers cannot use "." as an artifact_id and
accidentally operate on the whole artifacts tree.
oxoxDev
requested changes
May 28, 2026
Contributor
oxoxDev
left a comment
oxoxDev
left a comment
There was a problem hiding this comment.
Author: @graycyrus (
MEMBER— core team)
Substantial new subsystem — artifact persistence layer with proper structure, tests, and schema definitions. Core architecture is sound but one critical security issue + three majors need addressing before merge. coderabbit caught all four; surfacing here to gate the CHANGES_REQUESTED.
Blocker — verified against current code
🔴 src/openhuman/artifacts/store.rs:26-57 — validate_artifact_id allows "." → full artifacts-root deletion
Coderabbit's finding is correct and I traced it end-to-end:
validate_artifact_idonly rejects..,../,..\\—.passesdelete_artifact(workspace, ".")→root.join(".")=rootitselfassert_within_root(root, root.join("."))→starts_withcheck passes (PathBuf normalizes"./"to justroot)tokio::fs::remove_dir_all(&artifact_dir)wipes the entire artifacts tree
Reproducer: ai_delete_artifact(&config, ".") would delete every artifact in the workspace.
Required fix — reject . explicitly + simplify the .. check (the standalone-equal .. case is already covered by the new arm):
fn validate_artifact_id(id: &str) -> Result<(), String> {
if id.is_empty() {
return Err("[artifacts] artifact_id must not be empty".to_string());
}
if id == "." || id == ".." {
return Err(format!(
"[artifacts] artifact_id must not be a reserved path component: {id:?}"
));
}
if id.contains('/') { /* unchanged */ }
if id.contains('\\') { /* unchanged */ }
if id.starts_with("../") || id.starts_with("..\\") {
return Err(format!(
"[artifacts] artifact_id must not be a path traversal: {id:?}"
));
}
// rest unchanged
}Also add a regression test:
#[test]
fn validate_artifact_id_rejects_dot() {
assert!(validate_artifact_id(".").is_err());
}
#[tokio::test]
async fn delete_artifact_rejects_dot_does_not_wipe_root() {
let tmp = tempfile::tempdir().unwrap();
save_artifact_meta(tmp.path(), &dummy_meta("real-id")).await.unwrap();
assert!(delete_artifact(tmp.path(), ".").await.is_err());
// root + sibling artifact still exist
assert!(tmp.path().join("artifacts/real-id/meta.json").exists());
}Major — verified
🟠 src/openhuman/artifacts/ops.rs:63-69 — absolute_path derivation allows escape via meta.path
meta.path is joined directly without validation:
let absolute_path = config.workspace_dir.join("artifacts").join(&meta.path)If a stored ArtifactMeta.path is absolute or contains .. (a corrupted/malicious meta.json that survived disk inspection), absolute_path escapes the artifacts root and is returned to the caller. The caller may then pass it to file-read tools.
Coderabbit's suggested fix (reject is_absolute() + Component::ParentDir / RootDir / Prefix) is correct. Apply it.
🟠 src/openhuman/artifacts/schemas.rs:87-92 — get_artifact schema contract mismatch
Output schema advertises artifact as Ref("ArtifactMeta") but the doc-comment + code adds an absolute_path field. Discovery clients won't know about the extra field. Coderabbit's TypeSchema::Object shape fix is right.
🟠 .claude/memory.md:217 — security doc steers future domains away from security/policy.rs
The memory note advises future authors to "replicate this pattern" for new filesystem-backed domains. That conflicts with the documented policy that file I/O validation should route through src/openhuman/security/policy.rs::validate_path_within_root (verified at line 2030). The artifacts module reimplements containment checks locally instead of delegating — that's a defensible per-module choice, but the memory note shouldn't propagate it as the recommended pattern.
Coderabbit's doc rewrite (clarify the local pattern is domain-specific; new domains should use policy.rs entry points) is correct.
Stronger suggestion — consider whether this PR should also be migrated to use validate_path_within_root:
// replace assert_within_root with the policy-canonical entry point
let resolved = crate::openhuman::security::policy::validate_path_within_root(
&artifact_dir, &root
)?;validate_path_within_root canonicalizes both paths before starts_with, which closes the symlink-escape gap that assert_within_root (pure starts_with) doesn't catch. If <workspace>/artifacts/<id> is a symlink pointing outside the workspace, this PR's check passes; the canonical check would catch it. Worth at minimum a comment explaining why symlink resolution was deemed unnecessary if the local check stays.
Minor / nitpicks
src/openhuman/artifacts/store.rs:233—_assert_status_useddead-code helper feels like a smell; ifArtifactStatusis only referenced in tests, mark the type#[cfg_attr(not(test), allow(dead_code))]or move the test-only usage into a#[cfg(test)]block that uses it directly.src/openhuman/artifacts/store.rs:18—artifacts_rootcallscreate_dir_allon every invocation including read paths (list_artifacts,get_artifact). Mildly wasteful + races on first concurrent call. Splitting intoartifacts_root_create()(write paths) vsartifacts_root_existing()(read paths) would be cleaner.
Verified / looks good
- Module structure (
types/store/ops/schemas/mod) is clean and matches openhuman conventions - Pagination (
offset/limitwithMAX_LIMIT=200) follows established RPC patterns - Corrupt-meta skipping with
log::warn!is the right call (vs propagating) ArtifactKind::parse/ArtifactStatus::parsecorrectly fall back toOther/Pendingon unknown values- Test scaffolding present (
store_tests.rs,ops_tests.rs) - 1214 lines all-new with no deletions — clean greenfield
- coderabbit CHANGES_REQUESTED captures all the security issues
Recommendation
CHANGES_REQUESTED — primarily on the critical . deletion vector. Once that + the two security majors (ops absolute_path, memory doc) land, this is approve-ready.
M3gA-Mind
approved these changes
May 28, 2026
6 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
artifactsdomain (src/openhuman/artifacts/) withArtifactMetastruct,ArtifactKindenum (Presentation, Document, Image, Other), andArtifactStatusenum (Pending, Ready, Failed)<workspace>/artifacts/<uuid>/meta.json+ binary blobs with lazy directory creationai_list_artifacts(paginated),ai_get_artifact(metadata + absolute path),ai_delete_artifact(removes metadata + blob)Test plan
cargo checkpassescargo test -p openhuman --lib -- "artifacts"— 38/38 tests passcargo fmt --checkcleanChecklist
[artifacts]prefix on all operationsCloses #2776
Summary by CodeRabbit
New Features