[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M69926W8Q3Y2QW1T424HAC1B":3,"skill-same:01M69926W8Q3Y2QW1T424HAC1B":73},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":49,"commitSha":50,"skillDirectory":51,"forksCount":52,"repositoryHtmlUrl":53,"repositoryHomepageUrl":54,"ownerHtmlUrl":55,"ownerBlogUrl":56,"ownerDescription":57,"updatedAt":58,"install":59,"resources":72},"01M69926W8Q3Y2QW1T424HAC1B","mantis-reflect",null,"Extracts learnings from execution trajectories at the end of a Mantis loop. Use to parse agent conversations, extract successes, failures, and false assumptions, and append them to workspace\u002Flearnings.jsonl. Don't use for analyzing source code or writing patches.","google\u002Fmantis","google","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1342004?v=4","01M698VYEFDM0VJBZDYYJ2T83S","Google","01M6990SDGGFEAP21TJNS7HEG9",2778,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-reflect\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"mantis reflect","reflect","parse","writing","source code","source","for","at","conversations","jsonl","trajectories","false","end","learnings","use","append","loop","execution","analyzing","code","2026-11-10 10:12:30",false,"---\nname: mantis-reflect\ndescription: >-\n  Extracts learnings from execution trajectories at the end of a Mantis loop.\n  Use to parse agent conversations, extract successes, failures, and false assumptions, and append them to workspace\u002Flearnings.jsonl.\n  Don't use for analyzing source code or writing patches.\n---\n\n# Reflector (\u002Fmantis-reflect)\n\n## System Goal\n\nExecution Trajectory Analyst. Analyzes the sequence of thoughts, tool calls, and\nobservations (the \"trajectory\" or \"conversation\") of the other Mantis agents.\nExtracts valuable insights to prevent future agents from making the same\nmistakes.\n\n## Command Definition\n\n- **Command:** `\u002Fmantis-reflect`\n- **Description:** Parses execution trajectories from the current loop and\n  appends structured insights to `workspace\u002Flearnings.jsonl`.\n\n## Input\u002FOutput Contract\n\n- **Reads**:\n  - `workspace\u002F.mantis_state.json` (to track the current loop pass, and to read\n    `active_snapshot.snapshot_id` for provenance stamping — see Instructions\n    step 3). Read this file STATE-RELATIVE under `--state_root`; never derive\n    snapshot state by running live VCS.\n  - Subagent execution logs (`transcript.jsonl` files). The schema\n    `execution_log_entry` defined in `schema.json` is the normalized\n    representation. The orchestrator\u002Fadapter must normalize raw logs from\n    unsupported frameworks before passing them, or the reflector must parse\n    unsupported formats on a best-effort basis.\n  - **Locating Logs (harness-neutral):** The orchestrator SHOULD pass the list\n    of absolute file paths to the execution log files (e.g. `transcript.jsonl`)\n    for the subagents executed during this round; when provided, use these paths\n    directly. Do NOT hardcode any single framework's log layout. If no path list\n    was passed, resolve transcript paths from the ACTIVE harness's own\n    transcript convention — this is harness-specific and there are several. For\n    example, Antigravity stores them under\n    `\u003CappDataDir>\u002Fbrain\u002F\u003Cconversation_id>\u002F.system_generated\u002Flogs\u002Ftranscript.jsonl`;\n    other harnesses (e.g. Gemini CLI, the Google ADK, Claude Code) use different\n    layouts. Antigravity is ONE example among several, not the default. If,\n    after both routes, no readable transcript exists for a stage that ran this\n    round, do NOT abort and do NOT silently emit zero learnings — record a\n    missing-transcript insight per Instructions step 1.\n- **Writes**:\n  - Appends structured trajectory insights to `workspace\u002Flearnings.jsonl`.\n- **Preconditions**:\n  - Execution logs for the current round SHOULD exist and contain entries. If a\n    stage's log is missing, unreadable, empty, or yields zero parseable entries,\n    this is NOT a fatal error and NOT a reason to stop: continue with the other\n    stages and record the gap as a `trajectory_insight` (see Instructions step\n    1\\) so an absent log is never a silent zero-learnings result.\n- **Idempotency Guarantee**:\n  - Parses logs and filters already-recorded learnings to prevent duplicate\n    entries in `workspace\u002Flearnings.jsonl`. It should check existing lines in\n    `workspace\u002Flearnings.jsonl` to ensure it doesn't duplicate the same insight\n    if retried.\n  - When de-duplicating, compare on the semantic content (`target_entity` +\n    `insight` + `source_stage`) and treat `snapshot` as attached metadata, NOT\n    part of the identity, so a retry within the same pass does not\n    double-append. Missing-transcript insights (step 1) are de-duplicated the\n    same way.\n\n## Instructions\n\n### 0. Locator Resolution (Block A — FINDINGS-ONLY role)\n\n```\nLOCATOR RESOLUTION (before reading ANY target code or artifact):\n0. ROLE: If this skill NEVER reads target source (report, calibrate, reflect),\n   you are a FINDINGS-ONLY stage: skip steps 2-6; still read active_snapshot from\n   state for provenance\u002Fannotation; NEVER stop merely because a code root is unset.\n1. Determine CODE_ROOT, in this priority order:\n   a. If --target_root is passed on THIS invocation, CODE_ROOT = --target_root.\n      It is AUTHORITATIVE and OVERRIDES SNAPSHOT_ROOT and the state fallback\n      (used when a caller hands you a prepared tree, e.g. a patched shadow).\n   b. Else if --snapshot_root (or SNAPSHOT_ROOT) is passed, use it.\n   c. Else read state_root\u002Fworkspace\u002F.mantis_state.json (state_root from\n      --state_root if passed, else .\u002Fworkspace\u002F... relative to the current dir)\n      -> active_snapshot.root \u002F .snapshot_id \u002F .snapshot_pinned.\n   d. Else (no arg AND no readable active_snapshot): CODE_ROOT = current directory,\n      treat snapshot_pinned = false (MODE-OFF). Do NOT stop.\n2. SENTINEL CHECK (only if snapshot_pinned is true AND you did NOT take path 1a):\n   verify CODE_ROOT\u002F.mantis_snapshot_id exists and equals SNAPSHOT_ID. If missing\n   or different -> STOP \"snapshot sentinel mismatch\". (A --target_root tree (1a) is\n   deliberately mutated and is sentinel-EXEMPT.)\n3. PATH FIELDS:\n   - SNAPSHOT-RELATIVE (read under CODE_ROOT): code_paths entries; plan target_files\n     that are file paths. Strip ONLY a trailing \":\u003Cdigits>\". A code_paths entry\n     containing \":\u002F\u002F\" is a URL\u002Fendpoint, NOT a file read. A code_paths entry that is\n     NOT of the form \u003Cexisting-path>:\u003Cinteger> is a non-source LOCATOR\n     (symbol\u002Foffset\u002Fendpoint): only check that the artifact\u002Fsymbol exists; skip ALL\n     line-range and line-existence logic.\n   - STATE-RELATIVE (read\u002Fwrite under state_root\u002Fworkspace, NEVER prefix CODE_ROOT):\n     kb_references, repro_file_path, reattack_file_path, helper scripts, report\n     files, and all state\u002Ffindings JSON.\n4. Never WRITE under CODE_ROOT when snapshot_pinned is true. Any command that\n   compiles, generates, or writes artifacts MUST run in a PRIVATE SHADOW copy\n   (mktemp -d from CODE_ROOT), never with cwd=CODE_ROOT. Read-only inspection may\n   cd into CODE_ROOT.\n5. VCS-METADATA CARVE-OUT: history-log extraction and any VCS diff\u002Fblame command\n   run in the LIVE repository root (which still has .git\u002F.hg\u002F.repo), NOT CODE_ROOT\n   (the snapshot copy strips VCS metadata). Do NOT stop merely because CODE_ROOT\n   lacks .git\u002F.hg\u002F.repo.\n6. Every shell command uses ABSOLUTE paths and sets its own working directory on\n   that call. Do NOT assume the working directory persists between calls.\n```\n\nReflector is a FINDINGS-ONLY stage (same class as report and calibrate). Per\nBlock A step 0 it SKIPS steps 2-6 — it resolves NO CODE_ROOT, checks NO\n`.mantis_snapshot_id` sentinel, and applies NO snapshot-relative path logic —\nbut it STILL reads `active_snapshot` from state for provenance and NEVER stops\nbecause a code root is unset or unpinned. Reflector needs only `--state_root` to\nlocate `workspace\u002F.mantis_state.json`, `workspace\u002Flearnings.jsonl`, and the\ntranscript log files; all of these are STATE-RELATIVE (Block A step 3) and MUST\nNEVER be prefixed with CODE_ROOT. Reflector MUST NOT run any live VCS command\n(`git`\u002F`hg`\u002F`repo`); its only snapshot signal is `active_snapshot.snapshot_id`\nread from state.\n\nAnalyze the execution trajectories of all successfully executed subagents in the\nround (every stage that ran: history, summarize, architecture, threat-model,\nplan, researcher, dedupe, review, critic, reproduce, chain, patch, calibrate) to\ndistill what went right and what went wrong.\n\nExecute the reflection stage as follows:\n\n1. **Extract Trajectories (Token Optimization):**\n\n   - Do not attempt to read the entire, raw `transcript.jsonl` files natively\n     with `read_file`, as they can be massive and blow out your context window.\n\n   - Use the absolute log file paths passed to you by the orchestrator to access\n     the log files.\n\n   - Instead of reading the full files, use your bash\u002Fcommand execution tools to\n     parse and filter the logs. For example, write a short Python script or use\n     `jq`\u002F`grep` to extract key events (which should conform to the\n     `execution_log_entry` schema; if raw logs from different frameworks are\n     provided, parse them on a best-effort basis): tool error messages, final\n     agent summaries, instances where an agent \"gave up\", or messages indicating\n     a trust boundary assumption was incorrect.\n\n   - **Missing \u002F empty logs (never silent):** For each stage that ran this round\n     whose transcript is missing, unreadable, empty, or yields zero parseable\n     entries, append exactly one `trajectory_insight` with `action: \"add\"`,\n     `source_stage: \"mantis-reflect\"`,\n     `target_entity: \"\u003Cstage-name> transcript\"`, and an `insight` naming the\n     stage and the reason (e.g. \"no transcript file at \u003Cpath>\", \"empty log\", \"0\n     parseable entries\"). Do this BEFORE synthesizing insights, so even a fully\n     empty round produces auditable output instead of nothing.\n\n2. **Synthesize Insights:** Review the extracted events. Look for:\n\n   - **False Assumptions:** Did a researcher spend turns trying to exploit a\n     parameter, only to realize it was sanitized upstream in another file?\n   - **Tool Failures:** Did the reproducer fail consistently because of a\n     missing library in the sandbox?\n   - **Successful Strategies:** Did a patcher successfully fix a bug using a\n     specific idiomatic pattern that should be reused?\n\n3. **Append to the Inbox (`workspace\u002Flearnings.jsonl`):** For each distinct\n   insight, append a structured JSON object to `workspace\u002Flearnings.jsonl`.\n\n   ### Reflection Schema Format (`workspace\u002Flearnings.jsonl`)\n\n   ```json\n   {\"type\": \"trajectory_insight\", \"action\": \"add | update | remove\", \"target_entity\": \"[e.g., auth_module.py or sandbox_env]\", \"insight\": \"The researcher assumed input was unsanitized, but it is actually cleansed by the middleware. Do not attempt XSS on this parameter.\", \"source_stage\": \"mantis-researcher\", \"snapshot\": \"\u003Cactive_snapshot.snapshot_id from state; omit field entirely if unavailable>\"}\n   ```\n\n   **Snapshot provenance stamp (never live VCS):** Before writing, read\n   `active_snapshot.snapshot_id` from `workspace\u002F.mantis_state.json`\n   (STATE-RELATIVE, under `--state_root`). Set the OPTIONAL `snapshot` field on\n   EVERY emitted `trajectory_insight` (including the missing-transcript insights\n   from step 1) to that value, so each learning is attributable to the pass's\n   pinned snapshot. Do NOT run `git`\u002F`hg`\u002F`repo` or any live VCS command to\n   derive it. Backward-compat: if `active_snapshot` is absent, `snapshot` is\n   empty\u002Fnull, or state is unreadable, OMIT the `snapshot` field entirely and\n   proceed (degraded) — never stop and never fabricate an id. `snapshot` is an\n   OPTIONAL field: the `trajectory_insight` sub-schema in `schema.json` does not\n   set `additionalProperties: false`, so existing consumers accept it unchanged.\n\n   Ensure the file is appended to, not overwritten. When complete, notify the\n   user.\n","# Reflector (\u002Fmantis-reflect)\n\n## System Goal\n\nExecution Trajectory Analyst. Analyzes the sequence of thoughts, tool calls, and\nobservations (the \"trajectory\" or \"conversation\") of the other Mantis agents.\nExtracts valuable insights to prevent future agents from making the same\nmistakes.\n\n## Command Definition\n\n- **Command:** `\u002Fmantis-reflect`\n- **Description:** Parses execution trajectories from the current loop and\n  appends structured insights to `workspace\u002Flearnings.jsonl`.\n\n## Input\u002FOutput Contract\n\n- **Reads**:\n  - `workspace\u002F.mantis_state.json` (to track the current loop pass, and to read\n    `active_snapshot.snapshot_id` for provenance stamping — see Instructions\n    step 3). Read this file STATE-RELATIVE under `--state_root`; never derive\n    snapshot state by running live VCS.\n  - Subagent execution logs (`transcript.jsonl` files). The schema\n    `execution_log_entry` defined in `schema.json` is the normalized\n    representation. The orchestrator\u002Fadapter must normalize raw logs from\n    unsupported frameworks before passing them, or the reflector must parse\n    unsupported formats on a best-effort basis.\n  - **Locating Logs (harness-neutral):** The orchestrator SHOULD pass the list\n    of absolute file paths to the execution log files (e.g. `transcript.jsonl`)\n    for the subagents executed during this round; when provided, use these paths\n    directly. Do NOT hardcode any single framework's log layout. If no path list\n    was passed, resolve transcript paths from the ACTIVE harness's own\n    transcript convention — this is harness-specific and there are several. For\n    example, Antigravity stores them under\n    `\u003CappDataDir>\u002Fbrain\u002F\u003Cconversation_id>\u002F.system_generated\u002Flogs\u002Ftranscript.jsonl`;\n    other harnesses (e.g. Gemini CLI, the Google ADK, Claude Code) use different\n    layouts. Antigravity is ONE example among several, not the default. If,\n    after both routes, no readable transcript exists for a stage that ran this\n    round, do NOT abort and do NOT silently emit zero learnings — record a\n    missing-transcript insight per Instructions step 1.\n- **Writes**:\n  - Appends structured trajectory insights to `workspace\u002Flearnings.jsonl`.\n- **Preconditions**:\n  - Execution logs for the current round SHOULD exist and contain entries. If a\n    stage's log is missing, unreadable, empty, or yields zero parseable entries,\n    this is NOT a fatal error and NOT a reason to stop: continue with the other\n    stages and record the gap as a `trajectory_insight` (see Instructions step\n    1\\) so an absent log is never a silent zero-learnings result.\n- **Idempotency Guarantee**:\n  - Parses logs and filters already-recorded learnings to prevent duplicate\n    entries in `workspace\u002Flearnings.jsonl`. It should check existing lines in\n    `workspace\u002Flearnings.jsonl` to ensure it doesn't duplicate the same insight\n    if retried.\n  - When de-duplicating, compare on the semantic content (`target_entity` +\n    `insight` + `source_stage`) and treat `snapshot` as attached metadata, NOT\n    part of the identity, so a retry within the same pass does not\n    double-append. Missing-transcript insights (step 1) are de-duplicated the\n    same way.\n\n## Instructions\n\n### 0. Locator Resolution (Block A — FINDINGS-ONLY role)\n\n```\nLOCATOR RESOLUTION (before reading ANY target code or artifact):\n0. ROLE: If this skill NEVER reads target source (report, calibrate, reflect),\n   you are a FINDINGS-ONLY stage: skip steps 2-6; still read active_snapshot from\n   state for provenance\u002Fannotation; NEVER stop merely because a code root is unset.\n1. Determine CODE_ROOT, in this priority order:\n   a. If --target_root is passed on THIS invocation, CODE_ROOT = --target_root.\n      It is AUTHORITATIVE and OVERRIDES SNAPSHOT_ROOT and the state fallback\n      (used when a caller hands you a prepared tree, e.g. a patched shadow).\n   b. Else if --snapshot_root (or SNAPSHOT_ROOT) is passed, use it.\n   c. Else read state_root\u002Fworkspace\u002F.mantis_state.json (state_root from\n      --state_root if passed, else .\u002Fworkspace\u002F... relative to the current dir)\n      -> active_snapshot.root \u002F .snapshot_id \u002F .snapshot_pinned.\n   d. Else (no arg AND no readable active_snapshot): CODE_ROOT = current directory,\n      treat snapshot_pinned = false (MODE-OFF). Do NOT stop.\n2. SENTINEL CHECK (only if snapshot_pinned is true AND you did NOT take path 1a):\n   verify CODE_ROOT\u002F.mantis_snapshot_id exists and equals SNAPSHOT_ID. If missing\n   or different -> STOP \"snapshot sentinel mismatch\". (A --target_root tree (1a) is\n   deliberately mutated and is sentinel-EXEMPT.)\n3. PATH FIELDS:\n   - SNAPSHOT-RELATIVE (read under CODE_ROOT): code_paths entries; plan target_files\n     that are file paths. Strip ONLY a trailing \":\u003Cdigits>\". A code_paths entry\n     containing \":\u002F\u002F\" is a URL\u002Fendpoint, NOT a file read. A code_paths entry that is\n     NOT of the form \u003Cexisting-path>:\u003Cinteger> is a non-source LOCATOR\n     (symbol\u002Foffset\u002Fendpoint): only check that the artifact\u002Fsymbol exists; skip ALL\n     line-range and line-existence logic.\n   - STATE-RELATIVE (read\u002Fwrite under state_root\u002Fworkspace, NEVER prefix CODE_ROOT):\n     kb_references, repro_file_path, reattack_file_path, helper scripts, report\n     files, and all state\u002Ffindings JSON.\n4. Never WRITE under CODE_ROOT when snapshot_pinned is true. Any command that\n   compiles, generates, or writes artifacts MUST run in a PRIVATE SHADOW copy\n   (mktemp -d from CODE_ROOT), never with cwd=CODE_ROOT. Read-only inspection may\n   cd into CODE_ROOT.\n5. VCS-METADATA CARVE-OUT: history-log extraction and any VCS diff\u002Fblame command\n   run in the LIVE repository root (which still has .git\u002F.hg\u002F.repo), NOT CODE_ROOT\n   (the snapshot copy strips VCS metadata). Do NOT stop merely because CODE_ROOT\n   lacks .git\u002F.hg\u002F.repo.\n6. Every shell command uses ABSOLUTE paths and sets its own working directory on\n   that call. Do NOT assume the working directory persists between calls.\n```\n\nReflector is a FINDINGS-ONLY stage (same class as report and calibrate). Per\nBlock A step 0 it SKIPS steps 2-6 — it resolves NO CODE_ROOT, checks NO\n`.mantis_snapshot_id` sentinel, and applies NO snapshot-relative path logic —\nbut it STILL reads `active_snapshot` from state for provenance and NEVER stops\nbecause a code root is unset or unpinned. Reflector needs only `--state_root` to\nlocate `workspace\u002F.mantis_state.json`, `workspace\u002Flearnings.jsonl`, and the\ntranscript log files; all of these are STATE-RELATIVE (Block A step 3) and MUST\nNEVER be prefixed with CODE_ROOT. Reflector MUST NOT run any live VCS command\n(`git`\u002F`hg`\u002F`repo`); its only snapshot signal is `active_snapshot.snapshot_id`\nread from state.\n\nAnalyze the execution trajectories of all successfully executed subagents in the\nround (every stage that ran: history, summarize, architecture, threat-model,\nplan, researcher, dedupe, review, critic, reproduce, chain, patch, calibrate) to\ndistill what went right and what went wrong.\n\nExecute the reflection stage as follows:\n\n1. **Extract Trajectories (Token Optimization):**\n\n   - Do not attempt to read the entire, raw `transcript.jsonl` files natively\n     with `read_file`, as they can be massive and blow out your context window.\n\n   - Use the absolute log file paths passed to you by the orchestrator to access\n     the log files.\n\n   - Instead of reading the full files, use your bash\u002Fcommand execution tools to\n     parse and filter the logs. For example, write a short Python script or use\n     `jq`\u002F`grep` to extract key events (which should conform to the\n     `execution_log_entry` schema; if raw logs from different frameworks are\n     provided, parse them on a best-effort basis): tool error messages, final\n     agent summaries, instances where an agent \"gave up\", or messages indicating\n     a trust boundary assumption was incorrect.\n\n   - **Missing \u002F empty logs (never silent):** For each stage that ran this round\n     whose transcript is missing, unreadable, empty, or yields zero parseable\n     entries, append exactly one `trajectory_insight` with `action: \"add\"`,\n     `source_stage: \"mantis-reflect\"`,\n     `target_entity: \"\u003Cstage-name> transcript\"`, and an `insight` naming the\n     stage and the reason (e.g. \"no transcript file at \u003Cpath>\", \"empty log\", \"0\n     parseable entries\"). Do this BEFORE synthesizing insights, so even a fully\n     empty round produces auditable output instead of nothing.\n\n2. **Synthesize Insights:** Review the extracted events. Look for:\n\n   - **False Assumptions:** Did a researcher spend turns trying to exploit a\n     parameter, only to realize it was sanitized upstream in another file?\n   - **Tool Failures:** Did the reproducer fail consistently because of a\n     missing library in the sandbox?\n   - **Successful Strategies:** Did a patcher successfully fix a bug using a\n     specific idiomatic pattern that should be reused?\n\n3. **Append to the Inbox (`workspace\u002Flearnings.jsonl`):** For each distinct\n   insight, append a structured JSON object to `workspace\u002Flearnings.jsonl`.\n\n   ### Reflection Schema Format (`workspace\u002Flearnings.jsonl`)\n\n   ```json\n   {\"type\": \"trajectory_insight\", \"action\": \"add | update | remove\", \"target_entity\": \"[e.g., auth_module.py or sandbox_env]\", \"insight\": \"The researcher assumed input was unsanitized, but it is actually cleansed by the middleware. Do not attempt XSS on this parameter.\", \"source_stage\": \"mantis-researcher\", \"snapshot\": \"\u003Cactive_snapshot.snapshot_id from state; omit field entirely if unavailable>\"}\n   ```\n\n   **Snapshot provenance stamp (never live VCS):** Before writing, read\n   `active_snapshot.snapshot_id` from `workspace\u002F.mantis_state.json`\n   (STATE-RELATIVE, under `--state_root`). Set the OPTIONAL `snapshot` field on\n   EVERY emitted `trajectory_insight` (including the missing-transcript insights\n   from step 1) to that value, so each learning is attributable to the pass's\n   pinned snapshot. Do NOT run `git`\u002F`hg`\u002F`repo` or any live VCS command to\n   derive it. Backward-compat: if `active_snapshot` is absent, `snapshot` is\n   empty\u002Fnull, or state is unreadable, OMIT the `snapshot` field entirely and\n   proceed (degraded) — never stop and never fabricate an id. `snapshot` is an\n   OPTIONAL field: the `trajectory_insight` sub-schema in `schema.json` does not\n   set `additionalProperties: false`, so existing consumers accept it unchanged.\n\n   Ensure the file is appended to, not overwritten. When complete, notify the\n   user.\n",{"name":5,"description":7},"main","f3162c039ecd9c9d5d8bc404dd1373c7ca3152f7","skills\u002Fmantis-reflect",237,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis","https:\u002F\u002Fcloud.google.com\u002F","https:\u002F\u002Fgithub.com\u002Fgoogle","https:\u002F\u002Fopensource.google\u002F","Google ❤️ Open Source","2026-11-15 00:33:31",{"source":8,"skillName":5,"commands":60,"prompt":70,"zipAvailable":71},[61,64,67],{"tool":62,"command":63},"npx","npx skills add google\u002Fmantis --skill mantis-reflect",{"tool":65,"command":66},"bunx","bunx skills add google\u002Fmantis --skill mantis-reflect",{"tool":68,"command":69},"pnpm","pnpm dlx skills add google\u002Fmantis --skill mantis-reflect","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-reflect\u002FSKILL.md",true,[],[74,95,121,148,158,173],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":44},"01M6992XJHHV0R5N8E392EHRW3","mantis-threat-model","Synthesizes trust boundaries, attack surfaces, and attacker profiles into a living threat model. Use as Stage B of the Knowledge Base generation process, reading architecture and entity definitions from the KB. Don't use for analyzing source code or extracting raw learnings from JSONL files.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-threat-model\u002FSKILL.md",[],[81,82,83,84,85,43,42,86,87,88,89,90,37,91,92,93,94,33,29,28],"mantis threat model","threat model","model","threat","architecture","knowledge","kb","files","surfaces","trust","knowledge base","as","extracting","reading",{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":104,"publishedAt":44},"01M6992TR1BPA5HVF13QRM7EP0","mantis-summarize","Pre-processes the repository by generating security-focused summaries (mantis-summary.md) for each directory to make planning and research more efficient. Use when starting a review campaign to map the codebase before threat modeling and planning. Don't use for executing code reviews, writing test scripts, or patching code.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-summarize\u002FSKILL.md",[101],{"categoryNo":102,"categoryName":103,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000021","资料检索与研究",[105,106,27,107,108,109,110,43,111,112,113,84,114,115,116,117,118,119,30,120],"summarize","mantis summarize","directory","review","security","threat modeling","planning","repository","patching","campaign","map","modeling","scripts","make","reviews","when",{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":132,"publishedAt":44},"01M6992QB4DVMDR2AGT6BPSE3N","mantis-structural-index","Builds a content-addressed semantic-unit index from source code for structural context. Use when a pinned or live codebase is available and structural cross-reference data would improve research quality. Don't use for findings analysis, patching, or reporting.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-structural-index\u002FSKILL.md",[127],{"categoryNo":128,"categoryName":129,"taxonomyType":130,"taxonomyTypeDesc":131},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[133,134,135,136,43,137,138,139,140,113,141,142,143,38,144,145,146,147,30,120],"mantis structural index","index","structural","research","content","reporting","data","analysis","improve","quality","context","research quality","reference","live","unit",{"skillNo":149,"name":150,"nameZh":6,"description":151,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":152,"categories":153,"keywords":154,"publishedAt":44},"01M6992KRP2XYWEMV12P5M4GPE","mantis-review","Independently reviews findings and filters out false positives. Use when consolidated findings need validation against the actual source code. Don't use for reproducing crashes or patching code.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-review\u002FSKILL.md",[],[108,155,43,156,113,38,157,35,119,30,29,28,120],"mantis review","validation","filters",{"skillNo":159,"name":160,"nameZh":6,"description":161,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":162,"categories":163,"keywords":164,"publishedAt":44},"01M6992GKVGJ2NXEJWYFVCN01Q","mantis-researcher","Audits production source code files based on the strategy in workspace\u002Fplan.json. Use when a review plan exists and you need to perform static analysis and deep-dive reviews of targeted files. Don't use for planning, deduplicating, or writing patches.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-researcher\u002FSKILL.md",[],[165,166,27,108,167,43,111,140,168,169,38,88,170,171,119,30,29,172,28,120],"mantis researcher","researcher","static analysis","plan","strategy","deep","you","static",{"skillNo":174,"name":175,"nameZh":6,"description":176,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":177,"categories":178,"keywords":179,"publishedAt":44},"01M6992D5WXENF3TW68GM5WPHC","mantis-reproduce","Generates and runs crash reproducers to verify security flaws. Use when viable findings exist and you need to write and execute a script or payload to verify the crash. Don't use for code auditing or patching.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fmantis\u002Fblob\u002Fmain\u002Fskills\u002Fmantis-reproduce\u002FSKILL.md",[],[180,181,109,43,113,182,183,184,38,185,186,187,171,30,120],"mantis reproduce","reproduce","verify","write","execute","crash","payload","script"]