Skip to content

research: zeph-agent-tools and zeph-agent-context have no dedicated top-level spec #6637

Description

@bug-ops

Description

The specs quality audit in PR #6629 found that the `zeph-agent-tools` and `zeph-agent-context` crates (native tool loop, retry, confirmation, batch tool-result processing; system prompt rebuild, memory injection, summarization, compaction) have no dedicated top-level spec of their own — their behavior is covered piecemeal across specs `002` (agent-loop), `009` (orchestration), `033` (subagent-context-propagation), `044` (subagent-lifecycle), and `062` (context-adaptive-memory).

This is filed as research/backlog, not an active gap: per the sdd skill's own scaling guidance, not every crate needs a dedicated spec, and no concrete defect was found from the fragmentation. It's worth a scoping discussion on whether these two crates have grown enough independent surface area to warrant consolidation into their own specs, or whether the current piecemeal coverage remains adequate.

Expected Behavior

N/A — this is a scoping question, not a defined requirement yet.

Actual Behavior

N/A

Environment

Logs / Evidence

See PR #6629 description, "Flagged, not executed" section, item 7.

Metadata

Metadata

Assignees

No one assigned

    Labels

    P4Long-term / exploratorydocsDocumentationresearchResearch-driven improvement

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions