Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

test(protocol): complete MCP-Pi behavioral evidence and exact-candidate verification

Open
  • Is something wrong?
#1,985 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
25/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Active
Tech stack
go
Domain
backend-api-design, testing

Research direction

After #1981–#1984 resolve, read #1347 and map the Go ProfileAgent operations and agreed rules to existing tests in internal/mcp, internal/server, and plugin/pi/test. Add representative public interface vectors for uncovered rules, then verify artifact freshness in a clean LF controlled checkout and record the exact candidate SHA, runtime identities, commands, outcomes, and pending proof. Done means an explicit scope verdict that keeps parity, published pair, and installed host claims tied to their respective evidence.

Written by the indexing model from the issue text.

Description

status:needs-review type:feature
Existing Issue Search
  • I searched existing issues and will add evidence to a canonical tracker when applicable.
🔍 Observed Consequence

Inventory equality, generated artifacts and selected behavior tests do not establish full MCP/Pi semantic parity. #1347 still lacks complete applicable cross-runtime evidence and final exact-candidate verification.

💡 Desired Outcome

Maintainers can evaluate the agreed contract from a traceable behavioral evidence matrix and exact-candidate results without confusing mocked forwarding, real-server parity, installed-host compatibility or release qualification.

📦 Affected Area

MCP Server (tools, transport)

Release Context

main branch; remaining scope recorded in #1347. Planning slice, not fresh current-main reproduction, executed candidate checks or release compatibility evidence.

🔄 Workarounds and Constraints

No deferred substitute for each slice's own tests/docs, new runtime negotiation (#1350), release publication (#1349), unrelated transport rewrite, universal rollback or automatic replay.

📎 Additional Context

Parent acceptance tracker: #1347. Bounded delivery child, not a replacement for the umbrella.

Acceptance criteria
  • After preceding slices resolve, map current Go ProfileAgent operations and agreed rules to named existing tests. Add only missing representative behavioral vectors and identify remaining uncovered rules.
  • Retain MCP/HTTP compact-search/history vectors and Pi regressions. Reuse #1902/#1903 runtime/integration evidence where applicable rather than duplicating their qualification scope.
  • Cover missing agreed ownership, defaults, validation, exact errors and observable state effects through public interfaces against isolated data. This does not replace the tests/docs required in each capability delivery.
  • Distinguish pre-mutation rejection, post-mutation reload failure and unknown write acknowledgement. Preserve operation-specific guarantees; never assert universal rollback or replay writes automatically.
  • Verify artifact freshness in a clean LF-controlled checkout. Preserve the historical focused Go exit 1 reported in #1347 as failed evidence; investigate CRLF/LF before regenerating or alleging semantic drift.
  • Record exact candidate SHA and relevant runtime/package identities, commands, exit codes, passes, failures, skips and pending/unavailable proof. Earlier checkout/main audit results do not certify a later candidate.
  • Provide an explicit final scope verdict before proposing stronger artifact status or parent closure. Do not claim published-pair or installed-host compatibility without the corresponding evidence. Parent closure and stronger parity claims remain separate maintainer decisions.
Sequencing and dependencies

Hard dependency: complete the preceding four slices before final verification:

  1. #1981: feat(pi): expose mem_update find and replace under the Go agent contract
  2. #1982: feat(pi): expose mem_context older_history under the Go agent contract
  3. #1983: feat(protocol): reconcile projection documentation and current-project alias decision
  4. #1984: test(protocol): verify Pi input-schema projections against the Go agent contract
Relevant surfaces

Existing internal/mcp and internal/server parity vectors; plugin/pi/test forwarding/real-server fixtures; contract/projection documentation and candidate-specific evidence.

Existing-work boundaries

Open/closed duplicate searches covered #1347 references, mem_update/find, older_history, Pi/schema and parity. Final evidence slice under #1347; integrated #1348 alignment and #1856/#1857 search/history remain preserved. Runtime qualification #1902/#1903 supplies related evidence, not full semantic parity.

Publication boundary

Created at the maintainer's request to split #1347 into five reviewable deliveries. This publication authorizes no implementation, assignment, approval, commits, PRs, package publication or parent closure. CONTRIBUTING.md requires maintainer-accepted proportional evidence and a real assignee before implementation. Tests/docs belong with each behavior change, not only the final evidence issue.

Dominant language
Go
Stars
7.1k
Forks
729
Avg merge
9h 31m
Merged PRs (30d)
333

Getting set up

Open in Codespaces

Starts the project's dev container in your browser, under your own GitHub account.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from Gentleman-Programming/engram

All issues in Gentleman-Programming/engram

Similar issues

More Go issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.