Pod

Available as Markdown and JSON. Pod is also available over MCP.

Reported issues for Code Analyze MCP

Pod holds 5 of 5 GitHub reports that passed its relevance review. This can include external user reports, maintainer-confirmed bugs, and concrete feature gaps. Treat them as evidence to inspect, not a count of distinct defects.

Back to Code Analyze MCP.

Most discussed

feat(metrics): extend error_subtype coverage to exec_command and analyze_symbol

Context

From docs/audit/2026-09-11-param-description-experiments-audit.md (PR #1509), finding F4.

Problem

The JSONL metrics schema (docs/METRICS.md) records only error_type (invalid_params/parse/unknown) and a tool-specific error_subtype, which today is populated only for edit_replace. There is no free-text error message field. As a result, exec_command's 126 invalid_params errors (1.2% error rate over 11,275 calls) and analyze_symbol's 11 invalid_params errors…

Read the thread · 2026-09-11 · closed · 1 comment

[FEATURE] Native coder-handoff subcommand backed by SQLite

Summary

Add a native coder-handoff subcommand to the aptu-coder binary, backed by SQLite, so coding-pipeline handoff data (currently proposed as a standalone Python script in the coder skill's own repo, clouatre-labs/agentic-coder-skill) can eventually live in the same compiled, tested binary the pipeline already depends on for code analysis.

Context

Read the thread · 2026-09-07 · closed · 1 comment

[BUG] analyze_file intermittently fails structured-output schema validation (missing arg_count)

Summary

mcp__aptu-coder__analyze_file intermittently returns a structured-content payload that fails MCP schema validation on the client side, with the error Structured content does not match the tool's output schema: .../calls/N must have required property 'arg_count'. This was discovered via an audit of Claude Code session transcripts and the server's own metrics logs (~/.local/share/aptu-coder/metrics-*.jsonl), not via a local repro — filing so it's tracked and can be…

Read the thread · 2026-09-12 · closed · 0 comments

fix(pagination): clarify or validate empty/malformed cursor semantics

Context

From docs/audit/2026-09-11-param-description-experiments-audit.md (PR #1509), finding F3 / recommendation R3.

Problem

Agents repeatedly pass an empty or malformed cursor value (rather than omitting the parameter) to analyze_file, analyze_directory, and analyze_symbol, causing JSON-parse-EOF and base64-decode failures instead of the intended "no cursor" behavior.

Evidence

32 real errors observed across goose sessions: 20 analyze_file JSON-parse-EOF, 5…

Read the thread · 2026-09-11 · closed · 0 comments

[REFACTOR] Add schema-level maximum constraint to analyze_symbol's follow_depth

Summary

Add a JSON Schema maximum constraint to analyze_symbol's follow_depth parameter so invalid depths are schema-rejected instead of relying on prose guidance alone.

Motivation

Read the thread · 2026-09-10 · closed · 0 comments

Most recent

The remaining reports are on the project's issue tracker.