Individual skill

manage-tmcra-memory

Manage and troubleshoot TMCRA long-term memory through the bundled MCP tools and Codex device authorization. Use when a user explicitly asks to remember, recall, inspect, verify, or wait for memory; configure local Writer or organizer providers; asks why a memory was used; asks to persist an important project decision; or needs to connect or reauthorize the installed Codex plugin.

plugins/reshuibuduo/tmcra-plugin-codex/skills/manage-tmcra-memory/SKILL.md

Read pinned source

Source commit d780783eba4ecc7c8b9429f95c9362df95f07c61

What was checked

Matches the observed repository HEAD
Yes
Scanner rules are current
Yes
Repository scan complete
No
Skill scan complete
No
Observed behavior
Network or broad shell
Last checked
Nov 4, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings)
1,531
Repository last updated
Nov 6, 2026
License
Apache-2.0

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files
Read excerpt

# Manage TMCRA Memory

Automatic recall and capture are handled by lifecycle hooks. Use this skill for explicit memory operations and verification.

## Inspect the latest automatic recall

- Call `tmcra_last_recall` when the user asks which memories were injected for the latest answer, why TMCRA recalled something, or to show the evidence used before an answer.
- Use the default `view=latest_answer` after an answer has finished. This survives the new inspection prompt and returns the receipt promoted by that answer's `Stop` Hook.
- Use `view=current_prompt` only when the user explicitly asks about the prompt Codex is processing right now.
- Do not run a fresh search for either request. The tool reads the exact lifecycle-Hook receipt for the current project.
- Present global and project evidence separately, including their counts. Say clearly when the latest recall returned no evidence.
- The tool output is intentionally user-facing. Never supplement it with local receipt paths, scope names, tenant identifiers, database paths, query diagnostics, or raw lifecycle logs.
- Automatic recall remains silent during ordinary turns. Do not turn successful recall into a warning or interrupt every answer with a receipt.