Resume any chat
Skill adarshnagrikar14/resume-any-chat/.agents/skills/resume-any-chat
Use when the user wants to inspect, search, normalize, resume, hand off, or optionally import local Codex, Claude, or Antigravity CLI conversation history. This skill provides a local viewer, provider-specific parsers, native resume commands, a Codex handoff flow, and an explicit experimental Codex-history import path.From its SKILL.md
npx -y skills add adarshnagrikar14/resume-any-chat --skill resume-any-chatAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
3 things to look at
- skips confirmationTells the agent to proceed without asking first, 1 time: "do not ask for `provider + session id` before showing the actual list".
- 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
- runs commandsInstructs the agent to run 8 commands, including `node plugins/resume-any-chat/runtime/cli.js export` and 7 more.
SKILL.md
5.2 KB, ~1.2k tokens by cl100k_base, as published. Nobody here has run it
Resume Any Chat
Use this skill for local multi-provider chat history work across the current user's home directory:
Codexfrom.codex/sessionsClaudefrom.claude/projectsplus.claude/history.jsonlAntigravityfrom.gemini/antigravity-cli/brain/*/.system_generated/logs
Workflow
- Default to listing useful recent sessions directly in-chat.
Show recent
Codex,Claude, andAntigravitysessions first unless the user already named a provider/session. The default professional list format is:
- recent
10 - per item show:
idtitlelast message
- then show exactly these next actions:
Continue in CodexContinue NativeSummariseImport into Codex History (Experimental)
- if session
cwdand current workspace differ, say that briefly and suggest the safer option - stop there unless the user explicitly picks one of those actions
- Refresh/export the current index:
node plugins/resume-any-chat/runtime/cli.js export
- Useful commands:
node plugins/resume-any-chat/runtime/cli.js latest
node plugins/resume-any-chat/runtime/cli.js latest claude --limit 15
node plugins/resume-any-chat/runtime/cli.js session claude <session-id>
node plugins/resume-any-chat/runtime/cli.js handoff antigravity <session-id>
node plugins/resume-any-chat/runtime/cli.js import-history claude <session-id>
Command mapping:
show all historyshow historylist historyshow recentshow me all history
For any of the above, you must run:
node plugins/resume-any-chat/runtime/cli.js latest
If the user names a provider, you must run:
node plugins/resume-any-chat/runtime/cli.js latest <provider>
- Launch the local 3-tab viewer when the user wants to browse sessions visually:
node plugins/resume-any-chat/runtime/cli.js serve
- Prefer the safe built-in flows:
- Native resume:
codex resume <session-id>claude --resume <session-id>agy --conversation <session-id>
- Cross-provider continuation:
- generate a Codex handoff packet
- launch a fresh Codex session from that handoff
- Experimental import:
- only when the user explicitly chooses
Import into Codex History (Experimental) - import a foreign session into Codex history without replacing the safer default flows
- only when the user explicitly chooses
Rules
- Do not treat foreign-session import as the default continuation path.
- prefer native resume or Codex handoff first
- use experimental Codex-history import only after explicit user choice
- Do not read Antigravity token-monitor/auth cache unless the user explicitly asks for that risky path.
- Prefer
transcript_full.jsonlovertranscript.jsonlfor Antigravity when both exist. - If a provider session has no reliable
cwd, do not auto-launch; return the command and the missing-cwd warning. show session <provider> <id>should include recent turns by default.- default: show the overview plus last
5-8meaningful messages verbose: include more context, including thinking/tool lines when useful
- default: show the overview plus last
- If the user says
continue here, do not tell them to leave the thread first.- Run
handoff <provider> <session-id>without--launch - use the generated summary and handoff file as context
- continue the work in the current Codex thread
- do not stop at
If you want, I can...unless the next task is genuinely ambiguous
- Run
- Never auto-resume, auto-handoff, auto-summarise, or auto-open anything immediately after listing.
- listing is one step
- action happens only after the user explicitly chooses it
Import into Codex History (Experimental)is opt-in only.- add it as a selectable action
- do not use it unless the user explicitly chooses it
- Do not answer
show all historyby describing, linking, or summarizingstate.json.state.jsonis an internal cache/export artifact- mention it only if the user explicitly asks for raw JSON, export files, cache debugging, or file paths
- For
show all historyand similar list requests, return the directlatestoutput itself.- do not replace it with totals-only summaries
- do not ask for
provider + session idbefore showing the actual list
- If the prior session already contains a clear unfinished task, continue that task directly.
- Example: if the last unfinished ask was
show full repo structure, start doing that - only ask a follow-up if multiple next steps are equally plausible or if risk requires confirmation
- Example: if the last unfinished ask was
- The rule above applies only after the user has already chosen
Continue in Codexor equivalent. - Do not invent a different menu or a generic conversational close.
- follow the list-first flow above
- avoid endings like
If you want, I can...when the user already asked to continue
- Avoid rigid reply menus like
tell me one of these only.- accept natural requests such as
latest claude,show me more codex,continue this here,open the viewer,resume that claude one
- accept natural requests such as
What ships with it: 1 file
277 B alongside SKILL.md
agents/
- openai.yaml277 B
Gives 0 of the 12 instructions most hr recruiting skills give in ~1.2k tokens
Counted across 356 of the 357 authors here whose files we hold, read 2026-08-07
- Quantify achievements with specific metricsin 14 of 356, across 6 files
- Keep the resume under two pagesin 14 of 356, across 6 files
- Request the full job description if not providedin 12 of 356, across 4 files
- Extract keywords and prioritize job requirementsin 12 of 356, across 4 files
- Stop and ask for clarification if required inputs are missingin 12 of 356, across 5 files
- Map candidate experience to job requirementsin 11 of 356, across 3 files
- Ask if the user wants adjustmentsin 11 of 356, across 3 files
- Provide strengths and gap analysis after the resumein 10 of 356, across 2 files
- Request candidate background details if not providedin 10 of 356, across 2 files
- Format experience bullets as action verb plus resultin 10 of 356, across 2 files
- Ask for missing inputs before startingin 10 of 356, across 9 files
- Use exact job description terminologyin 9 of 356, across 1 file
Said here and by no other author read
- list recent sessions directly in-chat
- run latest for general history list requests
- show overview plus recent turns for a session
- continue an unfinished task directly
- Refresh the index before listing sessions
- Prefer native resume commands
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.