feat: hermes sessions import command

[iRonin]

Problem

hermes sessions export can export sessions to JSONL, but there is no corresponding import command. The exported JSONL is effectively one-way (backup/analysis only).

Proposed Solution

1. SessionDB.import_session() in hermes_state.py

• Insert session dict + messages into SQLite
• Collision: skip by default, --force overwrites
• Dangling parent_session_id: set to NULL
• Unknown columns stripped (future-proofing)
• JSON fields re-serialized automatically

2. CLI subcommand

hermes sessions import <file>           # import from JSONL
hermes sessions import -                # import from stdin
hermes sessions import backup.jsonl --force    # overwrite colliding IDs
hermes sessions import backup.jsonl --dry-run  # preview without writing
hermes sessions import backup.jsonl --source cli  # filter by source

3. Tests

• 12 unit tests for import_session()
• 10 CLI integration tests
• Round-trip: export -> import -> all messages preserved

[GottZ]

This was generated by AI during triage.

Summary

One PR, #8651, addresses #8650 by adding JSONL session import, collision handling, filtering, dry-run support, and tests. Its implementation matches the requested command at a high level, but the diff does not safely round-trip current exports because it targets an outdated schema, does not normalize structured message content through the current write path, and does not exercise the real CLI entry point.

Related pull requests

• feat: add sessions import command #8651 best fix — (+675/-0) — partial fix and best existing fix: feat: add sessions import command #8651 implements the requested import workflow, but its hard-coded schema-v6 allowlists omit current schema-v19 fields, raw insertion can fail for decoded multimodal content, and the CLI tests recreate a parser or call SessionDB directly instead of testing the actual command. This agrees with the keep_open review on feat: add sessions import command #8651, which requires a current-schema rewrite and real end-to-end CLI coverage.

Suggested consolidation

Keep #8651 open with a salvage path as the best existing fix: preserve its command shape, collision semantics, source filtering, dry-run behavior, and import test scenarios, while rewriting persistence around the current schema and shared normalization/write behavior and replacing the synthetic parser tests with end-to-end tests of the real CLI command. No duplicate PRs need closing.

Complex graph

flowchart LR
    classDef open fill:#dbeafe,stroke:#1d4ed8,color:#1e3a8a
    classDef merged fill:#dcfce7,stroke:#15803d,color:#14532d
    classDef closed fill:#e5e7eb,stroke:#6b7280,color:#1f2937
    classDef unverified fill:#f3f4f6,stroke:#9ca3af,color:#374151
    classDef best stroke-width:3px,stroke:#b45309
    classDef target stroke-width:3px,stroke:#4338ca
    I8650(["issue #8650 (open)"])
    P8651["PR #8651 (open)"]
    P8651 -->|best fix| I8650
    class I8650 open
    class I8650 target
    class P8651 open
    class P8651 best
    click I8650 "https://github.com/NousResearch/hermes-agent/issues/8650"
    click P8651 "https://github.com/NousResearch/hermes-agent/pull/8651"

Loading

Graph: solid arrow = fixes / best fix, dashed arrow = partial or unverified (see edge label); boxed group = PRs duplicating each other; amber border = best fix; indigo border = target; gray node = closed (state tag in the node label).

Cross-PR triage: Reviewed 1 pull request and 1 issue in this complex. Each diff was read against this issue; Assessment working set: 27 kB of PR diffs, 2 kB of issue/PR text, 1 kB of discussion (2 comments), 2 verify verdicts. verdicts reflect diff content, not PR titles. Part of an automated triage batch.