Add verified resource resolver before exposing external resources to agents
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 9
- Forks
- 17
- Avg merge
- 21h 36m
- Merged PRs (30d)
- 22
Description
Context
Context construction should not just retrieve useful information; it should also control which resources are allowed to enter the agent's working context. A model may hallucinate a plausible repo, package, tool, skill, or MCP server. If contextweaver blindly enriches that reference, the downstream agent may treat the wrong resource as legitimate.
Goal
Add a verified resource-resolution layer that converts ambiguous resource references into provenance-backed context cards.
Proposed scope
- Implement or prototype a resolver abstraction for:
- GitHub repositories
- package names
- MCP servers / tools
- URLs
- Emit a provenance card compatible with
weaver-speconce defined. - Support resolution outcomes:
- verified
- ambiguous
- unknown
- suspicious
- denied
- Add risk flags for:
- newly created or low-signal resources
- typo-like names
- owner mismatch
- no releases / no tags / no README
- install scripts or instructions asking for shell/network/filesystem access
- Make unsafe or ambiguous resources visible to the caller instead of silently injecting them into context.
Acceptance criteria
- Context cards can include requested identity, resolved identity, resolver evidence, and risk flags.
- Ambiguous or suspicious resources are not silently treated as trusted context.
- Tests cover at least one correct resolution, one ambiguous resolution, and one unverified/hallucinated resource.
- README includes a short example of safe resource resolution before agent execution.
Design principle
Context is part of the security boundary. The model should not be the authority for resource identity.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing the existing context construction flow and the README, then identify where external resource references enter agent context. Define the resolver outcomes and provenance-backed context-card fields described here, add tests for correct, ambiguous, and unverified resources, and document a safe resolution example in the README.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- github, python
- Domain
- ai, security
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100