Update to AiAkiv Memory
Snapshot Nov 6, 2026 · 12:10 UTC · version 2.3.0
Collection source: downloaded plugin package.
Instructions updated for aiakiv-links
Skill instructions — Now says: member** of the team (any role), together with the link list. A `caller_not_member` error means the user is not a member of the active team, so neither...
Observed in instructions or declared skills. Runtime behavior has not been tested.
Up to 3 plugins free · no card required.
All fields and original values.
Skill instructions
Oct 26, 2026 · 18:05 UTC
member**, even though the link contract list is owner-only. `links: []` with `links_unavailable_reason: "caller_not_owner"` and a `note` means the user is not the team owner, not that nothing is readable: use the public attaches th...
Nov 6, 2026 · 12:10 UTC
member** of the team (any role), together with the link list. A `caller_not_member` error means the user is not a member of the active team, so neither the links nor the public attaches are visible from it.
Compare saved observations
Download comparison JSONSKILL.md text changes
− removed · + added
--- before +++ after @@ -47,10 +47,10 @@ - A `usable` public attach works in the same four tools with that `link_id`, exactly like a link. Results carry `side: "remote"` the same way, and the same rule applies: keep them labelled as the public org's, not the user's. -- The section comes back to **every member**, even though the link contract - list is owner-only. `links: []` with `links_unavailable_reason: - "caller_not_owner"` and a `note` means the user is not the team owner, not - that nothing is readable: use the public attaches that are listed. +- The section comes back to **every member** of the team (any role), together + with the link list. A `caller_not_member` error means the user is not a + member of the active team, so neither the links nor the public attaches are + visible from it. - `blocked_reason: "attach_not_enabled"`, or a `__public:` id that is not listed, means the attach is off or the org is no longer public. Turning it on is a console action by the team owner; do not retry.
Full technical diff · 1 changed fields
changed /skill_md_contents
"---\nname: aiakiv-links\ndescription: >-\n Read a LINKED partner org's memory through the partner memory tools. Use when\n the user asks what a partner/linked org/team knows, wants to compare their\n memory with a partner's, mentions a link or partner org, or when a probe\n candidate carries side=remote. Covers choosing between search_partner_memory,\n find_partner_memory_connections and query_partner_memory_graph, reading results, and what each error reason\n means. Also covers public attaches (public memory attached to the user's\n team, listed under public_attaches with __public: ids).\n---\n\n# AiAkiv links (partner-org memory)\n\nA **link** is a read capability two orgs agreed to — not a project you switch\ninto. You pass `link_id` per call; your own scope is untouched. Everything is\nread-only: you can never write into a partner org.\n\n**Always start with `list_partner_links`.** It is local (works even when the\npartner is down) and tells you which links are `usable` before you spend a\nremote call. If the tools are absent from the tool list, the server has links\ndisabled — say so.\n\nLinks run **one way**, and the list holds the ones whose reading direction\npoints at the user's org — blocked ones included, with `usable` saying which\nanswer right now. Links their org **provides** (their team issued the invite, so\nthe link opens their memory to the partner) are left out entirely — nothing to\nread, nothing to fix; `provided_link_count` says how many exist, and they are\nmanaged in the console.\n\n**Call link tools one at a time.** Your org may have only one link call in\nflight at once. Two link tools issued in the same batch — the habit that serves\nyou well everywhere else — means the second returns `error: busy` without\nrunning, even on a completely idle server. Await each call before starting the\nnext; `list_partner_links` is local and does not count.\n\n## Public attaches (`public_attaches`)\n\n`list_partner_links` also returns `public_attaches`: public memory the user's\nteam attached to itself in the console (team \"Links\" page, \"Public attach\"\nsection). There is no contract and no partner on the other side. A public org\nis open to read anyway; the attach only adds name bridges (aliases) so walks\nand graph queries can cross into it. Each row has `kind: \"public\"`, a\n`link_id` of the form `__public:<org_id>`, `partner_name`, `usable`,\n`blocked_reason` and `alias_count`.\n\n- A `usable` public attach works in the same four tools with that `link_id`,\n exactly like a link. Results carry `side: \"remote\"` the same way, and the\n same rule applies: keep them labelled as the public org's, not the user's.\n- The section comes back to **every member**, even though the link contract\n list is owner-only. `links: []` with `links_unavailable_reason:\n \"caller_not_owner\"` and a `note` means the user is not the team owner, not\n that nothing is readable: use the public attaches that are listed.\n- `blocked_reason: \"attach_not_enabled\"`, or a `__public:` id that is not\n listed, means the attach is off or the org is no longer public. Turning it\n on is a console action by the team owner; do not retry.\n- `alias_count: 0` means searches and bodies still work but walks and graph\n queries stay home-only (no bridge). Say so instead of reporting absence.\n\n## Which tool for which question\n\n| Question shape | Tool |\n|---|---|\n| \"What does the partner know about X?\" — search THEIR memory directly | `search_partner_memory(sentence, link_id)` — entry happens on their side; results are their events only |\n| \"From what WE know, what connects to their side?\" — walk from your context across shared ground | `find_partner_memory_connections` — entry in YOUR org, expansion crosses through **aliased entities** (entities the link declared \"same thing on both sides\") |\n| Explicit relations across both orgs (shared entities, vector distance, ranked bridges) | `query_partner_memory_graph(link_id, query, params)` — same language as `query_memory_graph`; see the `aiakiv-graph-query` skill |\n| Read one partner event's body | `get_partner_memory_content(link_id, event_id)` — paged; `event_id` comes from the other tools' results |\n\n## Reading results\n\n- Partner events carry `side: \"remote\"` (or a `link_id`). **Keep them\n labelled as the partner's** when you reason or summarize — do not present\n partner knowledge as the user's own, and do not re-save partner content\n into the user's memory unless the user explicitly asks.\n- `find_partner_memory_connections` never claims something is *absent* on the partner side —\n entry is home-only, so the partner is only reached through bridges. Absence\n of remote candidates means \"no bridge found\", not \"they don't know\".\n- Aliased bridge entities carry `df_home` / `df_remote` (how common the\n entity is on each side). Asymmetry is a signal: \"1 here, 109 there\" means a\n passing mention for the user is a central topic for the partner — often the\n most interesting finding in the response.\n- `found: false` from `get_partner_memory_content` means the link does not\n surface that event — deliberately the same answer whether it is absent or\n out of scope. Don't retry; don't speculate which.\n\n## When a call returns `status: \"error\"`\n\nThe `reason` names what to do — an error here **never** means the user's own\nmemory failed:\n\n- `no_links`, `link_not_found`, `link_not_active`, `link_expired`,\n `org_not_party`, `principal_not_active` — this link cannot be read (or this\n direction is closed). Show `list_partner_links` output; fixing it is an\n owner/console action, not a retry.\n - One case has nothing to fix: a `link_id` that `list_partner_links` does not\n list at all (the user may have copied it from the console's \"links we\n provide\" group) is one their org **provides**, and its refusal is\n `principal_not_active`. Say that it opens their memory outward rather than\n sending them to fix a direction.\n- `attach_not_enabled` — a `__public:` id whose attach is off, or whose org is\n no longer public (see \"Public attaches\" above). Owner/console action.\n- `contract_version_too_old` — both owners must re-consent;\n `list_partner_links` → `reconsent` shows who is missing.\n- `budget_exceeded` — the remote row/time budget ran out mid-walk. Narrow the\n ask (smaller `top_n`/`k`/`LIMIT`, tighter relation params) instead of\n repeating it.\n- `busy` — a concurrency limit on **your** server (never the partner's). The\n `detail` says which of two: the per-org limit, which is almost always\n self-inflicted (two link calls issued together) — re-issue the refused one\n once the other finishes, it cost nothing and did not run; or the server-wide\n limit, shared with other tenants, where a short wait is the right response.\n Either way it is not a partner problem.\n- `remote_failure`, `internal` — infrastructure failed; try later.\n\nResults can also be `partial` (budget hit mid-walk): what came back is valid,\njust incomplete — say so instead of treating it as the full picture.\n"
"---\nname: aiakiv-links\ndescription: >-\n Read a LINKED partner org's memory through the partner memory tools. Use when\n the user asks what a partner/linked org/team knows, wants to compare their\n memory with a partner's, mentions a link or partner org, or when a probe\n candidate carries side=remote. Covers choosing between search_partner_memory,\n find_partner_memory_connections and query_partner_memory_graph, reading results, and what each error reason\n means. Also covers public attaches (public memory attached to the user's\n team, listed under public_attaches with __public: ids).\n---\n\n# AiAkiv links (partner-org memory)\n\nA **link** is a read capability two orgs agreed to — not a project you switch\ninto. You pass `link_id` per call; your own scope is untouched. Everything is\nread-only: you can never write into a partner org.\n\n**Always start with `list_partner_links`.** It is local (works even when the\npartner is down) and tells you which links are `usable` before you spend a\nremote call. If the tools are absent from the tool list, the server has links\ndisabled — say so.\n\nLinks run **one way**, and the list holds the ones whose reading direction\npoints at the user's org — blocked ones included, with `usable` saying which\nanswer right now. Links their org **provides** (their team issued the invite, so\nthe link opens their memory to the partner) are left out entirely — nothing to\nread, nothing to fix; `provided_link_count` says how many exist, and they are\nmanaged in the console.\n\n**Call link tools one at a time.** Your org may have only one link call in\nflight at once. Two link tools issued in the same batch — the habit that serves\nyou well everywhere else — means the second returns `error: busy` without\nrunning, even on a completely idle server. Await each call before starting the\nnext; `list_partner_links` is local and does not count.\n\n## Public attaches (`public_attaches`)\n\n`list_partner_links` also returns `public_attaches`: public memory the user's\nteam attached to itself in the console (team \"Links\" page, \"Public attach\"\nsection). There is no contract and no partner on the other side. A public org\nis open to read anyway; the attach only adds name bridges (aliases) so walks\nand graph queries can cross into it. Each row has `kind: \"public\"`, a\n`link_id` of the form `__public:<org_id>`, `partner_name`, `usable`,\n`blocked_reason` and `alias_count`.\n\n- A `usable` public attach works in the same four tools with that `link_id`,\n exactly like a link. Results carry `side: \"remote\"` the same way, and the\n same rule applies: keep them labelled as the public org's, not the user's.\n- The section comes back to **every member** of the team (any role), together\n with the link list. A `caller_not_member` error means the user is not a\n member of the active team, so neither the links nor the public attaches are\n visible from it.\n- `blocked_reason: \"attach_not_enabled\"`, or a `__public:` id that is not\n listed, means the attach is off or the org is no longer public. Turning it\n on is a console action by the team owner; do not retry.\n- `alias_count: 0` means searches and bodies still work but walks and graph\n queries stay home-only (no bridge). Say so instead of reporting absence.\n\n## Which tool for which question\n\n| Question shape | Tool |\n|---|---|\n| \"What does the partner know about X?\" — search THEIR memory directly | `search_partner_memory(sentence, link_id)` — entry happens on their side; results are their events only |\n| \"From what WE know, what connects to their side?\" — walk from your context across shared ground | `find_partner_memory_connections` — entry in YOUR org, expansion crosses through **aliased entities** (entities the link declared \"same thing on both sides\") |\n| Explicit relations across both orgs (shared entities, vector distance, ranked bridges) | `query_partner_memory_graph(link_id, query, params)` — same language as `query_memory_graph`; see the `aiakiv-graph-query` skill |\n| Read one partner event's body | `get_partner_memory_content(link_id, event_id)` — paged; `event_id` comes from the other tools' results |\n\n## Reading results\n\n- Partner events carry `side: \"remote\"` (or a `link_id`). **Keep them\n labelled as the partner's** when you reason or summarize — do not present\n partner knowledge as the user's own, and do not re-save partner content\n into the user's memory unless the user explicitly asks.\n- `find_partner_memory_connections` never claims something is *absent* on the partner side —\n entry is home-only, so the partner is only reached through bridges. Absence\n of remote candidates means \"no bridge found\", not \"they don't know\".\n- Aliased bridge entities carry `df_home` / `df_remote` (how common the\n entity is on each side). Asymmetry is a signal: \"1 here, 109 there\" means a\n passing mention for the user is a central topic for the partner — often the\n most interesting finding in the response.\n- `found: false` from `get_partner_memory_content` means the link does not\n surface that event — deliberately the same answer whether it is absent or\n out of scope. Don't retry; don't speculate which.\n\n## When a call returns `status: \"error\"`\n\nThe `reason` names what to do — an error here **never** means the user's own\nmemory failed:\n\n- `no_links`, `link_not_found`, `link_not_active`, `link_expired`,\n `org_not_party`, `principal_not_active` — this link cannot be read (or this\n direction is closed). Show `list_partner_links` output; fixing it is an\n owner/console action, not a retry.\n - One case has nothing to fix: a `link_id` that `list_partner_links` does not\n list at all (the user may have copied it from the console's \"links we\n provide\" group) is one their org **provides**, and its refusal is\n `principal_not_active`. Say that it opens their memory outward rather than\n sending them to fix a direction.\n- `attach_not_enabled` — a `__public:` id whose attach is off, or whose org is\n no longer public (see \"Public attaches\" above). Owner/console action.\n- `contract_version_too_old` — both owners must re-consent;\n `list_partner_links` → `reconsent` shows who is missing.\n- `budget_exceeded` — the remote row/time budget ran out mid-walk. Narrow the\n ask (smaller `top_n`/`k`/`LIMIT`, tighter relation params) instead of\n repeating it.\n- `busy` — a concurrency limit on **your** server (never the partner's). The\n `detail` says which of two: the per-org limit, which is almost always\n self-inflicted (two link calls issued together) — re-issue the refused one\n once the other finishes, it cost nothing and did not run; or the server-wide\n limit, shared with other tenants, where a short wait is the right response.\n Either way it is not a partner problem.\n- `remote_failure`, `internal` — infrastructure failed; try later.\n\nResults can also be `partial` (budget hit mid-walk): what came back is valid,\njust incomplete — say so instead of treating it as the full picture.\n"
Full snapshot data
{
"description": "Read a LINKED partner org's memory through the partner memory tools. Use when the user asks what a partner/linked org/team knows, wants to compare their memory with a partner's, mentions a link or partner org, or when a probe candidate carries side=remote. Covers choosing between search_partner_memory, find_partner_memory_connections and query_partner_memory_graph, reading results, and what each error reason means. Also covers public attaches (public memory attached to the user's team, listed under public_attaches with __public: ids).",
"included_files": [],
"name": "aiakiv-links",
"skill_md_contents": "---\nname: aiakiv-links\ndescription: >-\n Read a LINKED partner org's memory through the partner memory tools. Use when\n the user asks what a partner/linked org/team knows, wants to compare their\n memory with a partner's, mentions a link or partner org, or when a probe\n candidate carries side=remote. Covers choosing between search_partner_memory,\n find_partner_memory_connections and query_partner_memory_graph, reading results, and what each error reason\n means. Also covers public attaches (public memory attached to the user's\n team, listed under public_attaches with __public: ids).\n---\n\n# AiAkiv links (partner-org memory)\n\nA **link** is a read capability two orgs agreed to — not a project you switch\ninto. You pass `link_id` per call; your own scope is untouched. Everything is\nread-only: you can never write into a partner org.\n\n**Always start with `list_partner_links`.** It is local (works even when the\npartner is down) and tells you which links are `usable` before you spend a\nremote call. If the tools are absent from the tool list, the server has links\ndisabled — say so.\n\nLinks run **one way**, and the list holds the ones whose reading direction\npoints at the user's org — blocked ones included, with `usable` saying which\nanswer right now. Links their org **provides** (their team issued the invite, so\nthe link opens their memory to the partner) are left out entirely — nothing to\nread, nothing to fix; `provided_link_count` says how many exist, and they are\nmanaged in the console.\n\n**Call link tools one at a time.** Your org may have only one link call in\nflight at once. Two link tools issued in the same batch — the habit that serves\nyou well everywhere else — means the second returns `error: busy` without\nrunning, even on a completely idle server. Await each call before starting the\nnext; `list_partner_links` is local and does not count.\n\n## Public attaches (`public_attaches`)\n\n`list_partner_links` also returns `public_attaches`: public memory the user's\nteam attached to itself in the console (team \"Links\" page, \"Public attach\"\nsection). There is no contract and no partner on the other side. A public org\nis open to read anyway; the attach only adds name bridges (aliases) so walks\nand graph queries can cross into it. Each row has `kind: \"public\"`, a\n`link_id` of the form `__public:<org_id>`, `partner_name`, `usable`,\n`blocked_reason` and `alias_count`.\n\n- A `usable` public attach works in the same four tools with that `link_id`,\n exactly like a link. Results carry `side: \"remote\"` the same way, and the\n same rule applies: keep them labelled as the public org's, not the user's.\n- The section comes back to **every member** of the team (any role), together\n with the link list. A `caller_not_member` error means the user is not a\n member of the active team, so neither the links nor the public attaches are\n visible from it.\n- `blocked_reason: \"attach_not_enabled\"`, or a `__public:` id that is not\n listed, means the attach is off or the org is no longer public. Turning it\n on is a console action by the team owner; do not retry.\n- `alias_count: 0` means searches and bodies still work but walks and graph\n queries stay home-only (no bridge). Say so instead of reporting absence.\n\n## Which tool for which question\n\n| Question shape | Tool |\n|---|---|\n| \"What does the partner know about X?\" — search THEIR memory directly | `search_partner_memory(sentence, link_id)` — entry happens on their side; results are their events only |\n| \"From what WE know, what connects to their side?\" — walk from your context across shared ground | `find_partner_memory_connections` — entry in YOUR org, expansion crosses through **aliased entities** (entities the link declared \"same thing on both sides\") |\n| Explicit relations across both orgs (shared entities, vector distance, ranked bridges) | `query_partner_memory_graph(link_id, query, params)` — same language as `query_memory_graph`; see the `aiakiv-graph-query` skill |\n| Read one partner event's body | `get_partner_memory_content(link_id, event_id)` — paged; `event_id` comes from the other tools' results |\n\n## Reading results\n\n- Partner events carry `side: \"remote\"` (or a `link_id`). **Keep them\n labelled as the partner's** when you reason or summarize — do not present\n partner knowledge as the user's own, and do not re-save partner content\n into the user's memory unless the user explicitly asks.\n- `find_partner_memory_connections` never claims something is *absent* on the partner side —\n entry is home-only, so the partner is only reached through bridges. Absence\n of remote candidates means \"no bridge found\", not \"they don't know\".\n- Aliased bridge entities carry `df_home` / `df_remote` (how common the\n entity is on each side). Asymmetry is a signal: \"1 here, 109 there\" means a\n passing mention for the user is a central topic for the partner — often the\n most interesting finding in the response.\n- `found: false` from `get_partner_memory_content` means the link does not\n surface that event — deliberately the same answer whether it is absent or\n out of scope. Don't retry; don't speculate which.\n\n## When a call returns `status: \"error\"`\n\nThe `reason` names what to do — an error here **never** means the user's own\nmemory failed:\n\n- `no_links`, `link_not_found`, `link_not_active`, `link_expired`,\n `org_not_party`, `principal_not_active` — this link cannot be read (or this\n direction is closed). Show `list_partner_links` output; fixing it is an\n owner/console action, not a retry.\n - One case has nothing to fix: a `link_id` that `list_partner_links` does not\n list at all (the user may have copied it from the console's \"links we\n provide\" group) is one their org **provides**, and its refusal is\n `principal_not_active`. Say that it opens their memory outward rather than\n sending them to fix a direction.\n- `attach_not_enabled` — a `__public:` id whose attach is off, or whose org is\n no longer public (see \"Public attaches\" above). Owner/console action.\n- `contract_version_too_old` — both owners must re-consent;\n `list_partner_links` → `reconsent` shows who is missing.\n- `budget_exceeded` — the remote row/time budget ran out mid-walk. Narrow the\n ask (smaller `top_n`/`k`/`LIMIT`, tighter relation params) instead of\n repeating it.\n- `busy` — a concurrency limit on **your** server (never the partner's). The\n `detail` says which of two: the per-org limit, which is almost always\n self-inflicted (two link calls issued together) — re-issue the refused one\n once the other finishes, it cost nothing and did not run; or the server-wide\n limit, shared with other tenants, where a short wait is the right response.\n Either way it is not a partner problem.\n- `remote_failure`, `internal` — infrastructure failed; try later.\n\nResults can also be `partial` (budget hit mid-walk): what came back is valid,\njust incomplete — say so instead of treating it as the full picture.\n"
}SHA-256 of public snapshot: b7718daae33e7bd8028ebaff7aa6f9004d13f132290928a679d01620fd04991f